See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/372921665

Advances in Artificial Intelligence for Image Processing: Techniques,

Applications, and Optimization

Chapter · June 2023

DOI: 10.4018/978-1-6684-8618-4.ch006

CITATIONS READS
93 4,209

3 authors, including:

Samapth Boopathi Binay Kumar Pandey

Muthayammal Engineering College 161 PUBLICATIONS 8,189 CITATIONS
231 PUBLICATIONS 9,798 CITATIONS
SEE PROFILE
SEE PROFILE

All content following this page was uploaded by Samapth Boopathi on 05 August 2023.

The user has requested enhancement of the downloaded file.

 73

Chapter 6
Advances in Artificial
Intelligence for Image
Processing:
Techniques, Applications,
and Optimization

Sampath Boopathi
https://orcid.org/0000-0002-2065-6539
Muthayammal Engineering College, India

Binay Kumar Pandey

https://orcid.org/0000-0002-4041-1213
College of Technology, Govind Ballabh Pant University of Agriculture and Technology, India

Digvijay Pandey
https://orcid.org/0000-0003-0353-174X
Department of Technical Education, India

ABSTRACT
AI has had a substantial influence on image processing, allowing cutting-edge methods and uses.
The foundations of image processing are covered in this chapter, along with representation, formats,
enhancement methods, and filtering. It digs into methods for machine learning, neural networks, opti-
mization strategies, digital watermarking, picture security, cloud computing, image augmentation, and
data pretreatment methods. The impact of cloud computing on platforms, performance, privacy, and
security are also covered. The chapter’s consideration of future trends and applications emphasises
the substantial contributions that AI has made to image processing as well as the ethical and societal
ramifications of this technology.

DOI: 10.4018/978-1-6684-8618-4.ch006

Copyright © 2023, IGI Global. Copying or distributing in print or electronic forms without written permission of IGI Global is prohibited.
 
Advances in Artificial Intelligence for Image Processing

INTRODUCTION

AI has a huge influence on image processing by providing cutting-edge techniques and applications.
The basics of image processing, representation, formats, enhancement techniques, filtering, machine
learning, neural networks, optimization techniques, digital watermarking, picture security, cloud com-
puting, image augmentation, and data pretreatment are covered in this chapter. It also talks about how
platforms, performance, privacy, and security are affected by cloud computing. Future developments
and applications will demonstrate the important advances AI has achieved in image processing while
simultaneously tackling moral and societal issues(Anitha et al., 2023; Reddy et al., 2023). An impor-
tant development in image processing is deep learning, a subfield of AI that makes use of artificial
neural networks. Convolutional Neural Networks (CNNs) have demonstrated effectiveness in a variety
of tasks, including segmentation, object recognition, and image categorization. These deep learning
architectures achieve excellent generalisation and accuracy levels by extracting key features from raw
visual input(Alam et al., 2022).
A generator and discriminator neural network are combined to create GANs, which are AI image
processing methods. They have important uses in a variety of fields, including healthcare, where they
enhance diagnostic accuracy, early sickness detection, and customised treatment plans. In dermatology,
pathology, and radiology, AI has also excelled in identifying abnormalities, diseases, and medical pro-
fessionals. Robots are now able to recognise and evaluate visual information like humans thanks to AI
in image processing(Janardhana, Anushkannan, et al., 2023; Jeevanantham et al., 2023; Selvakumar et
al., 2023). AI algorithms are used by autonomous vehicles for safe navigation, and surveillance systems
monitor activity and boost safety. The entertainment industry has benefited from AI’s expansion of
creativity by making content production, video editing, and special effects possible. In order to increase
the efficacy and efficiency of algorithms, optimization is essential in AI image processing. Optimizing
the training and inference procedures is crucial as the complexity of deep learning models rises. Deep
neural network training is accelerated by methods including parallel computing, distributed learning,
and hardware acceleration, enabling the deployment of real-time applications on devices with limited
resources(Letourneau-Guillon et al., 2020; Malik et al., 2018).
The utilisation of computational resources like memory and energy efficiently is necessary for im-
proved AI algorithms for processing images. Model compression techniques like pruning, quantization,
and knowledge distillation reduce the size of deep learning models and the amount of compute required
without significantly reducing performance. The deployment of AI-powered image processing apps on
a variety of platforms, including edge devices and cloud-based infrastructure, is made possible by these
optimization approaches(Khokhar et al., 2015).
AI has had a big influence on image processing thanks to cutting-edge methods like deep learning,
CNNs, and GANs. These methods allow for the creation of hierarchical representations from data, which
makes them advantageous for usage in fields including surveillance, self-driving cars, healthcare, and
transportation. AI image processing techniques have improved, allowing for real-time applications and
deployment on limited devices(A. Mohanty, Jothi, et al., 2023; Rahamathunnisa et al., 2023; Samikannu
et al., 2023). This article will highlight recent developments and their effects on the area of image process-
ing by concentrating on particular approaches, applications, and optimization techniques. The creation of
intelligent systems that can carry out activities that need human intelligence is known as artificial intel-
ligence (AI). It combines computer vision, natural language processing, and machine learning. Innova-
tions in deep learning architectures, neural networks, and machine learning algorithms have propelled the

74

Advances in Artificial Intelligence for Image Processing

development of AI(Baduge et al., 2022). AI has uses in a number of industries, including entertainment,
banking, healthcare, and transportation. This summary offers a strong framework for comprehending
the importance of AI in image processing. In fields like computer vision, medical imaging, and digital
media, image processing is essential(Boopathi, Arigela, et al., 2023; A. Mohanty, Venkateswaran, et
al., 2023; Senthil et al., 2023). AI significantly contributes to the automation of processes like photo
processing, producing results that are quicker and more accurate(Anitha et al., 2023; Babu et al., 2023;
Boopathi, Arigela, et al., 2023; Jeevanantham et al., 2023; Subha et al., 2023). Accuracy enhancements,
efficiency improvements, and the capacity for handling big datasets are all advantages. Autonomous
cars, facial recognition software, and image analysis in medicine are examples of practical uses. These
examples show how much AI has affected image processing(Sun et al., 2019).
The chapter focuses on artificial intelligence (AI) methods and their usage in image processing,
including computer vision methods, deep learning frameworks, and machine learning algorithms. It
emphasises current developments and new patterns, including explainable AI, transfer learning, and
GANs. Potential difficulties are also discussed in the chapter, including data privacy, interpretability,
and moral issues(Babu et al., 2023; Boopathi, Khare, et al., 2023; Vennila et al., 2023). It also looks at
potential study areas and future approaches, giving readers a comprehensive overview of AI in image
processing that covers both present state-of-the-art methods and foreseeable futures. This thorough
analysis of AI in image processing gives readers a good knowledge of its background, goals, and focus
areas. They are better equipped to explore AI-based image processing and its many applications with
this expertise(Abduljabbar et al., 2019).

FUNDAMENTALS OF IMAGE PROCESSING

Basics of Digital Images

Digital pictures, which are discrete representations of visual information in pixels, require an understand-
ing of image processing. The number of pixels in the grid determines resolution, with higher resolutions
offering more information. Grayscale and colour digital pictures are often used, with grayscale images
having a single channel for intensity or brightness levels and colour photos having three channels for
colour intensity. These channels are combined to provide a depiction in full colour(Bagheri et al., 2019).

Image Representation and Formats

There are several file formats that may be used to encode and save images, each having its own features
and methods of compression. The popular file types include JPEG, PNG, BMP, and GIF. JPEG is a
lossy format for natural photos that achieves excellent compression ratios by choosing which information
to keep and which to delete. PNG is a lossless format used for transparency or fine detail preservation
that supports both grayscale and colour pictures(Wang et al., 2020). In Windows systems, BMP is a
straightforward uncompressed picture format that stores data as a series of pixels without compression.
Using LZW compression, the GIF format is compressed and ideal for basic visuals and colour schemes.
By storing several photos in a single file, it facilitates animation.

75
 
Advances in Artificial Intelligence for Image Processing

Image Enhancement Techniques

By adjusting a picture’s attributes including brightness, contrast, sharpness, and colour balance, image
enhancement techniques seek to increase the visual quality of a given image. These methods draw atten-
tion to crucial details, reduce noise, and improve the image’s overall attractiveness. A typical method for
improving contrast that makes use of the whole dynamic range of the image is histogram equalisation.
By emphasising high-frequency components, image sharpening methods like unsharp masking and La-
placian sharpening filters improve edges and details. By adjusting colour balance and saturation, colour
correction ensures that colours seem natural and consistent on various devices and in a variety of light-
ing situations(Berg et al., 2019). Image enhancement requires noise reduction, especially when there is
little light or when the sensor sensitivity is high. Gaussian and median filters, for instance, smooth out
noise while keeping crucial visual information.

Image Filtering and Restoration

For photos to be of higher quality and be useful for analysis or display, artefacts, blur, and distortions
must be removed using image filtering and restoration procedures. By convolutional filtering a picture
with a preset kernel, linear filters, such as Gaussian and mean filters, blur and smooth images. Non-
linear filters that maintain edges and fine features while lowering noise include median and bilateral
filters. Deblurring techniques estimate and reverse the blurring process to restore sharpness and clarity.
They are used to treat problems like motion blur or defocus blur. Effective image processing requires a
thorough understanding of digital picture principles, representation formats, and image augmentation
and filtering methods. These fundamental ideas offer a strong framework for investigating sophisticated
methods and applications in AI-driven image processing(Kan, 2017).

MACHINE LEARNING FOR IMAGE ANALYSIS

Introduction to Machine Learning

A critical area of artificial intelligence called machine learning (ML) creates algorithms and models
that can recognise patterns and take wise judgments without explicit programming. A model is trained
on a dataset that has been labelled, where input data (pictures) is linked to matching output labels (clas-
sifications or annotations). In order to generalise and make predictions on data that has not yet been
observed, the model learns underlying patterns and relationships in the training data(Harikaran et al.,
2023; Koshariya et al., 2023; Subha et al., 2023; Vanitha et al., 2023). ML algorithms may be divided
into three categories: reinforcement learning, unsupervised learning, and supervised learning(Madabhushi
& Lee, 2016).

Supervised Learning Algorithms for Image Classification

Models are trained using labelled examples in the supervised learning paradigm of machine learning,
which is utilised for image classification problems. SVMs, decision trees, and random forests are a few
examples of algorithms that learn decision boundaries in feature space to divide various classes. Con-

76

Advances in Artificial Intelligence for Image Processing

volutional neural networks (CNNs), in particular, have transformed picture categorization by process-
ing grid-like input data and detecting regional patterns as well as global structures. On difficult image
classification tasks, CNN architectures like AlexNet, VGGNet, and ResNet have shown state-of-the-art
results(Abduljabbar et al., 2019).

Unsupervised Learning Techniques for Image Clustering

In unlabeled training data, latent patterns or structures are uncovered using unsupervised learning ap-
proaches. Unsupervised learning methods are frequently employed in image analysis to cluster comparable
pictures based on visual similarity. Data is divided into groups using clustering techniques like k-means,
hierarchical clustering, and Gaussian mixture models, where photos inside a cluster are more similar to
one another than those in other clusters(Boopathi, 2023a; Janardhana, Singh, et al., 2023; Kavitha et al.,
2023; Sathish et al., 2023). Without depending on predetermined labels or annotations, these methods
facilitate the exploration and organisation of enormous picture collections.

Deep Learning Approaches for Image Recognition

Picture identification problems including object detection, image segmentation, and facial recognition
have been transformed by deep learning. In these challenging challenges, deep neural networks have
excelled by learning hierarchical representations from unprocessed visual input. R-CNN, YOLO, and SD
are a few object identification algorithms that can locate and categorise various things in a picture. Fully
Convolutional Networks (FCNs) and U-Net architectures are used in image segmentation techniques
to divide pictures into semantically significant sections(Boopathi, Siva Kumar, et al., 2023; Boopathi,
Venkatesan, et al., 2023; Gowri et al., 2023; Yupapin et al., 2023). Face recognition software learns
discriminative representations of faces for precise face detection, recognition, and attribute analysis.
Deep learning models are used to identify and validate persons based on facial traits. Learning similarity
metrics for face recognition is aided by methods such as Siamese networks and deep metric learning.

Transfer Learning in Image Processing

Transfer learning is a method that uses pretraining on one task or dataset to transfer information to an-
other task or dataset that is linked to the first. Due to the availability of extensive pretraining datasets
like ImageNet, it is essential in image processing. Deep learning models that have already been trained,
such as those on ImageNet, are capable of learning detailed visual representations from tagged photos
and may be used as a base for a variety of image processing applications. Transfer learning enables
faster convergence and enhanced performance on target tasks by reusing and optimising these models.
Feature extraction, where the convolutional layers of the pretrained model are employed as fixed feature
extractors, and fine-tuning, where both convolutional and classifier layers are tweaked on the target
task, are two ways to implement transfer learning. As a result, the model may take use of the general
information gathered during pretraining while also adapting to the target task’s unique properties(Berg
et al., 2019; Kan, 2017).
Finally, by enabling image categorization, grouping, identification, and other difficult tasks, machine
learning techniques have significantly enhanced image analysis tasks. These techniques consist of deep

77
 
Advances in Artificial Intelligence for Image Processing

learning, unsupervised learning, and supervised learning. Transfer learning has improved performance
on specific image processing applications by utilising pretrained models.

NEURAL NETWORKS IN IMAGE PROCESSING

The applications of neural networks in image processing are shown in Figure 1.

Basics of Neural Networks

Modern image processing and analysis need neural networks, which were inspired by the structure
and operation of the human brain. These computational models are made up of artificial neurons that
are linked to one another and carry out calculations and generate results. The artificial neuron, which
performs a nonlinear transformation on a weighted sum of inputs, is the fundamental component of a
neural network(Boopathi, 2023b; Harikaran et al., 2023; Jeevanantham et al., 2023; Rahamathunnisa et
al., 2023). With the help of training, neural networks may learn intricate mappings between inputs and
outputs, empowering them to anticipate the future and carry out tasks like image analysis(Baduge et al.,
2022; Madabhushi & Lee, 2016; Sun et al., 2019).

Convolutional Neural Networks (CNN) for Image Analysis

By using spatial linkages and local patterns in pictures, CNNs revolutionise image analysis. They are
made up of convolutional layers, pooling layers, and fully linked layers and are designed for input data
with a grid-like structure. Convolutional layers capture regional patterns like edges, corners, and textures
by using filters or kernels to extract features at various spatial positions.
By downsampling the features’ spatial dimensions, pooling layers aid in the extraction of the most
pertinent data while lowering the computing burden. Max and average pooling are two common pooling
procedures. Each layer’s neurons are linked by fully connected layers, which allow a network to learn
high-level representations and generate predictions. Large labelled datasets are used to train CNNs, while
backpropagation and gradient descent are used to maximise performance. Using the learnt network,
tasks like image

Recurrent Neural Networks (RNN) for Image Sequence Analysis

Recurrent neural networks (RNNs) are useful for applications like video analysis, captioning, and picture
descriptions because they are excellent for evaluating visual sequences and time-series data. RNNs may
preserve data from earlier inputs thanks to connections that create loops. In order to capture temporal
relationships and context in sequential data, the recurrent unit maintains an internal state that is modified
based on input and prior state. RNNs have the ability to process a sequence of pictures and create output
at each stage or to process the full sequence and then provide a final output. For the vanishing gradient
issue and long-term dependencies, popular alternatives include Long Short-Term Memory (LSTM) and
Gated Recurrent Unit (GRU)(Egmont-Petersen et al., 2002; Zhao et al., 2015).

78

Advances in Artificial Intelligence for Image Processing

Figure 1. Neural networks in image processing

Generative Adversarial Networks (GANs) for Image Generation

Neural networks called Generative Adversarial Networks (GANs) are employed to create brand-new
pictures or alter old ones. They are made up of a discriminator network and a generator network that can
discriminate between created and genuine pictures while learning to create realistic images. While the
discriminator seeks to accurately identify pictures, the generator and discriminator networks compete to
create indistinguishable images during training(Boopathi, 2022a, 2022b; A. Mohanty, Jothi, et al., 2023;
Samikannu et al., 2023). The generator strives to produce more realistic pictures as a result of this com-
petitive process. In picture creation tasks such creating realistic faces, original artwork, high-resolution
photos, image-to-image translation, and style transfer, GANs have shown amazing results(Mehdy et al.,
2017).

Reinforcement Learning for Image-Based Decision Making

In order to optimise reward signals, agents might make decisions or conduct actions in an environment
using the reinforcement learning (RL) paradigm. It is used in situations where agents must make decisions
based on pictures or video frames in order to comprehend and interact with their surroundings. These
inputs are processed by convolutional neural networks, which subsequently extract pertinent information
that can be utilised to decide or control activities. Successful applications of this strategy include robotic
control, autonomous driving, and game play(Ranganath et al., 1995). With CNNs excelling at extracting
spatial information from pictures, RNNs for evaluating image sequences, and Generative Adversarial
Networks (GANs) revolutionising image production, neural networks have considerably enhanced im-
age processing and analysis. Reinforcement Learning (RL) promotes innovation in the area of image
processing by enabling image-based decision-making in dynamic contexts.

79
 
Advances in Artificial Intelligence for Image Processing

IMAGE ANALYTICS AND OPTIMIZATION

Image Analytics and Optimization processes are illustrated in Figure 2.

Figure 2. Image analytics and optimization

Object Detection and Recognition

Identifying and categorising things of interest within a picture is the focus of the critical image analyt-
ics tasks of object detection and recognition. In order to acquire discriminative features from labelled
training data, machine learning algorithms like CNNs are frequently utilised. These characteristics may
subsequently be used to find and identify objects in fresh photos. Region-based Convolutional Neural
Networks (R-CNN), Faster R-CNN, and You Only Look Once are a few examples of object identification
frameworks (YOLO). Assigning class labels to the objects in an image is the process of object recogni-
tion. Deep learning models, such as CNNs, are very good at learning and identifying objects from visual
data because they can capture both low-level visual features and high-level semantic representations,
leading to robust and accurate object recognition(Letourneau-Guillon et al., 2020; Mikołajczyk et al.,
2018; Naranjo-Torres et al., 2020).

Image Segmentation and Clustering

Techniques for picture segmentation and clustering divide images into useful sections based on visual
similarity. With the use of methods like thresholding, region-growing, and graph-based segmentation,
pictures are divided into coherent sections. Semantic segmentation has demonstrated promising results
using deep learning techniques like FCNs and U-Net. Image clustering uses methods like k-means, hi-

80

Advances in Artificial Intelligence for Image Processing

erarchical clustering, and spectral clustering to group similar photos together without the need of labels.
Images can be grouped according to meaningful representations that can be learned using unsupervised
learning approaches, such as deep clustering algorithms(Li et al., 2018).

Feature Extraction and Dimensionality Reduction

The most useful and discriminative characteristics of pictures are captured using feature extraction and
dimensionality reduction algorithms, which also try to simplify the data. Raw picture data is converted
into representative features, such as texture descriptors, colour histograms, or SIFT features, through the
process of feature extraction. Using both low-level and high-level characteristics, deep learning techniques
like CNNs learn hierarchical representations from unstructured data. Techniques for reducing dimen-
sionality while keeping crucial information include Principal Component Analysis (PCA), t-distributed
Stochastic Neighbor Embedding (t-SNE), and autoencoders. These methods get rid of elements that are
unnecessary or redundant, increase computational effectiveness, and make it easier to visualise large
amounts of data(Chen et al., 2022).

Content-Based Image Retrieval

Rather than using textual annotations or metadata, Content-Based Image Retrieval (CBIR) is a technique
for finding and retrieving photographs based on their visual content. CBIR systems compare and match
query pictures with database images by examining visual characteristics including colour, texture, and
form(Boopathi, 2019, 2021). They compare feature representations and obtain visually related photos
using feature extraction methods like deep learning models or manually created descriptors, as well as
similarity measurements like Euclidean distance or cosine similarity. Search engines, picture database
management, and recommendation systems are just a few of the CBIR’s many uses(Anderson et al.,
2019; Grupac & L\uaz\uaroiu, 2022; Voyatzis et al., 1998).

Optimization Techniques in Image Processing

For image processing to be effective and produce the required results, optimization approaches are
crucial. Sharpening, denoising, and contrast enhancement are examples of enhancement methods that
try to maximise objective functions assessing picture quality. In image registration, where several pic-
tures are aligned to create composites or examine temporal or spatial changes, optimization techniques
are also applied. Based on similarity metrics like mutual information or cross-correlation, registration
algorithms optimise the transformation parameters(Boopathi, 2022c; Chakravarthi et al., 2022; Kannan
et al., 2022). For inverse issues like picture deblurring or reconstruction from sparse data, optimization
is crucial. Finding the most likely option that fulfils restrictions is made easier by iterative methods and
variational approaches. For extracting useful information, enhancing image quality, and facilitating ef-
fective retrieval, image analytics and optimization approaches are essential. The breakthroughs in image
processing and analysis include object identification, segmentation, feature extraction, content-based
picture retrieval, and optimization, with applications in computer vision, medical imaging, remote sens-
ing, and multimedia(Anderson et al., 2019; Grupac & L\uaz\uaroiu, 2022).

81
 
Advances in Artificial Intelligence for Image Processing

DIGITAL WATERMARKING AND IMAGE PROTECTION

Introduction to Digital Watermarking

Digital watermarking incorporates strong and undetectable information into digital media for functions
including content authentication, copyright protection, and tamper detection. Making embedded infor-
mation resistant to image processing procedures while yet being invisible to human observers is the aim.
Authorized parties must be able to recognise and remove the watermark using the proper methods(Voyatzis
et al., 1998; Yuan & Hao, 2020).

Techniques for Image Watermarking

There are a variety of image watermarking techniques, each having advantages and
disadvantages(Mohanarathinam et al., 2020).

• Least Significant Bit (LSB) and Spread Spectrum watermarking are examples of spatial domain
watermarking, which includes embedding a watermark directly into an image’s spatial domain by
changing pixel values.
• Using DFT or DWT, a watermark is embedded in transformed coefficients when an image is
converted into a new domain, such as frequency. These methods can withstand picture alteration
better.
• Statistical watermarking alters the distribution of the picture and provides resilience against as-
saults by embedding watermarks like histograms or distributions in the image’s attributes.
• Blind watermarking techniques are helpful when the original picture is unavailable because they
extract watermarks from watermarked photos without using the original, unwatermarked image
as a reference.

Robustness and Security in Image Watermarking

In picture watermarking, reliability and security are crucial. Strong watermarking techniques make
sure the watermark is still able to be seen and extracted despite malicious or accidental attacks. Using
encryption, digital signatures, and authentication codes, security focuses on preventing illegal removal
or manipulation. Collusion assaults, in which many watermarked copies work together to erase the wa-
termark, should also be addressed via secure systems(Arnold et al., 2003).

Copyright Protection and Authentication

For the copyright protection and verification of digital photographs, digital watermarking is essential.
Copyright owners can establish ownership and prevent unlawful use by including a distinctive watermark.
In the event of infringement, watermarks might include information such as copyright notices, author
information, or unique identifiers. Watermark extraction and comparison with the original reference
are the two steps that watermark authentication systems take to confirm the legitimacy and integrity
of a picture. These methods protect intellectual property rights, guarantee the accuracy of the material,
and allow for the safe distribution and use of digital photographs(S. P. Mohanty, 1999; Xuehua, 2010).

82

Advances in Artificial Intelligence for Image Processing

CLOUD COMPUTING FOR IMAGE PROCESSING

Introduction to Cloud Computing

By enabling scalable, on-demand access to computer resources through the internet, cloud computing
revolutionises data processing and storage. For image processing workloads demanding computing
power and storage capacity, it provides a versatile and affordable option. It is appropriate for a variety
of image processing applications because the cloud architecture, which is offloaded to distant servers,
provides a wide range of services, including virtual machines, storage, and specialised tools(Guo et al.,
2010; Yan & Huang, 2014).

Cloud-Based Image Processing Platforms

With pre-built algorithms, libraries, and frameworks, cloud-based image processing platforms offer tools
and services for image analysis and modification while streamlining the creation and deployment of ap-
plications. AWS, Google Cloud Platform, and Microsoft Azure are examples of well-known platforms. By
offering APIs and SDKs for developers to include image processing capabilities into their applications,
these platforms make it possible to upload, process, and store massive numbers of photos efficiently
while requiring less infrastructure setup and upkeep(Ferzli & Khalife, 2011; Yan & Huang, 2014).

Scalability and Performance Considerations

For activities involving image processing, cloud computing offers outstanding scalability, enabling dy-
namic resource allocation in response to demand. When processing huge datasets or during peak periods,
features like load balancing and auto-scaling enable effective completion. Network latency, data transfer
rates, and cloud infrastructure capabilities are all factors that affect performance. Cloud areas that are
near together geographically reduce latency and speed up processing. The efficiency of image processing
jobs may be improved by choosing cloud instances with enough processing capacity and by leveraging
distributed processing frameworks like Apache Spark or Hadoop(Kagadis et al., 2013; Xu et al., 2020).

Privacy and Security in Cloud-Based Image Processing

When processing private and sensitive photographic data in the cloud, privacy and security are essential.
The security measures used by cloud service providers range from encryption to access control systems
to routine audits. While data is being sent or stored, it is protected using encryption techniques like TLS
and data encryption at rest, while access control technologies like RBAC and IAM let users specify
precise access permissions. However, users are responsible for managing their own security procedures,
which may include defining access limits, upgrading frameworks and software, and using secure cod-
ing techniques. Protecting the privacy of picture data requires knowing the terms and conditions of the
cloud service provider and making sure that all legal requirements are met. To sum up, cloud computing
provides an effective and adaptable platform for image processing activities, but privacy and security
issues must constantly be taken into account to safeguard sensitive data and guarantee compliance with
privacy laws.

83
 
Advances in Artificial Intelligence for Image Processing

IMAGE AUGMENTATION AND DATA PREPROCESSING

Data Augmentation Techniques for Image Data

By adding modifications to existing photos, the data augmentation approach in image processing and
computer vision artificially increases the training dataset. This broadens variety and variability and en-
hances the resilience and generalisation of trained models(Tang et al., 2020; Tasci et al., 2021). Image
data may be enhanced using a variety of methods, such as:
Through rotation, translation, scaling, flipping, and shearing operations, geometric transformations
change the physical characteristics of a picture without changing its semantic information.
Brightness, contrast, saturation, and channel shifting adjustments are used in colour transformations
to change the colour attributes of a picture.
By modelling fluctuations using Gaussian or salt-and-pepper noise, noise augmentation increases a
model’s tolerance to real-world variances.
Images can take on different viewpoints and scales thanks to cropping and resizing processes, which
remove picture portions and change spatial resolution.
A broad and representative dataset for machine learning models is produced using augmentation
techniques, which supplement the training dataset with several versions of the original pictures.

Image Pre-Processing for Deep Learning

• By standardising and normalising input pictures, image preprocessing improves the performance
and training appropriateness of deep learning models. Typical preprocessing procedures include:
• By guaranteeing that all input photos have the same proportions, rescaling images to a set resolu-
tion assures uniform dimensions and streamlines the training process.
• By ensuring uniform pixel values across pictures, normalisation eliminates bias and improves
model learning by eradicating disparities in pixel ranges.
• By removing the mean pixel values from each pixel, mean subtraction decreases illumination vari-
ability and centres data around zero.
• Before being employed in deep learning models, images are transformed to tensors or arrays to
provide efficient computation and representation in formats like JPEG or PNG.

Data Imbalance and Sampling Techniques

When various classes in a dataset have uneven samples, biassed models result, which is known as data
imbalance. Different sampling approaches might be used during training to increase performance on
minority classes to overcome this issue(Minh et al., 2018).

• Through random duplication or more sophisticated methods like the Synthetic Minority Over-
sampling Technique (SMOTE), which creates synthetic samples based on existing minority sam-
ples, oversampling evens out class distribution by repeating minority cases.
• By deleting samples at random or using strategies like cluster-based undersampling, undersam-
pling evens out the distribution of the classes by lowering the number of instances from the domi-
nant class.

84

Advances in Artificial Intelligence for Image Processing

• When training a model, class weighting applies varying weights to samples from various classes,
emphasising on minority samples for more sensitive representation.
• By representing each class in a balanced training set in accordance with its original distribution,
stratified sampling preserves original class ratios and balances training sets.

Techniques for enhancing data address concerns with data imbalance, allowing models to success-
fully learn from all classes and enhance performance on minority classes. Rescaling, normalisation, and
mean subtraction are examples of image preprocessing techniques that guarantee standardised input for
deep learning models. Techniques include oversampling, undersampling, class weighting, and strati-
fied sampling reduce data imbalance, ensuring that picture data is properly enhanced and prepared for
reliable machine learning models.

REAL-TIME CASE STUDY: PACKING IN THE INDUSTRY

For items to be safely wrapped for storage, delivery, or sale, the packing process is essential in sectors
like manufacturing and logistics. Efficiency, accuracy, and cost reductions have all increased thanks to
AI and image processing tools. Automated packaging systems, including conveyor belts or robotic arms,
are designed to place items within containers appropriately, making the most of available space and
reducing risk of damage while being transported(Egmont-Petersen et al., 2002; S. P. Mohanty, 1999;
Xuehua, 2010).
In order to create an ideal packaging strategy that reduces empty space and avoids collisions, AI
algorithms evaluate 3D point cloud data to identify product size, shape, and orientation. As new goods
or container space changes, these algorithms may be dynamically changed. The ideal packing arrange-
ment is determined by advanced optimization algorithms that take into account the size, weight, fragility,
and industry-specific limitations of the product. Packing systems may achieve high efficiency, lower
packaging waste, and boost total output by merging AI and optimization methodologies.
By examining container photos to find flaws, irregularities, and missing products, artificial intelligence
(AI) assists in quality control throughout the packing process. This makes it possible for fast feedback
and remedial measures, ensuring that only containers that have been correctly packed are sent for ad-
ditional processing or transportation(Boopathi, Khare, et al., 2023; Kumara et al., 2023; A. Mohanty,
Venkateswaran, et al., 2023; Vennila et al., 2023).
The packaging business may profit greatly from AI and image processing thanks to increased pro-
ductivity, better space use, and higher-quality products. Industries may obtain effective, precise, and
affordable packaging solutions with the use of cutting-edge algorithms and real-time data analysis,
increasing production and customer satisfaction. This real-world case study shows how AI and image
processing may be used to improve product quality while streamlining the packing process and lowering
manual labour and human error.

Real-Time Case Study: AI for Image Processing in the Medical Field

Medical picture analysis and interpretation have been greatly enhanced by AI and image processing
approaches. The use of AI algorithms to analyse different medical pictures, including X-rays, CT scans,
MRI scans, ultrasound images, and histopathology slides, is highlighted in this real-time case study.

85
 
Advances in Artificial Intelligence for Image Processing

These pictures include important diagnostic information, but their proper and effective interpretation
might be difficult. Convolutional neural networks (CNNs) and generative adversarial networks (GANs),
which have been trained on sizable datasets of labelled medical images to learn patterns, features, and
abnormalities associated with various diseases or conditions, are two deep learning techniques com-
monly used to implement AI algorithms(Abduljabbar et al., 2019; Berg et al., 2019; Sun et al., 2019;
Tasci et al., 2021).
For tasks like picture segmentation, object identification, classification, and anomaly detection, AI
models are trained. In medical pictures, they can identify malignant tumours and segment anatomical
parts for surgical planning. With the use of ground truth annotations from qualified radiologists or pa-
thologists, these models are polished and verified.
AI has completely changed how medical image processing is done, leading to increased precision,
effectiveness, speedier diagnosis, and better identification of minor problems. With the help of these
algorithms, radiologists and pathologists may spot probable irregularities and use the information to
diagnose and detect diseases. Additionally, real-time analysis of medical pictures speeds up prompt
actions, especially in urgent situations, and decreases patient wait times. Additionally, AI systems are
able to spot tiny patterns or anomalies, increasing diagnostic sensitivity, enabling the early diagnosis
of diseases and improving patient outcomes. Medical imaging and other patient data may be analysed
by AI algorithms to offer individualised therapy suggestions that can be used to create treatment pro-
grammes that are specifically catered to each patient(Anitha et al., 2023; Babu et al., 2023; Boopathi,
Arigela, et al., 2023; Jeevanantham et al., 2023; Subha et al., 2023). Another advantage of AI algorithms
is augmented decision support, which offers extra information and suggestions based on medical picture
analysis, lowering diagnostic mistakes and boosting confidence in treatment choices.
Validation, regulatory compliance, and interaction with clinical procedures are all necessary for AI
used for image processing in the medical sector. Deep learning algorithms have the capacity to analyse
and interpret medical pictures, as shown by successful case studies and current research. Accuracy,
effectiveness, and tailored treatment are all improved by training AI models on massive datasets. This
leads to enhanced decision support for healthcare workers, quicker treatment planning, and increased
diagnostic capabilities, all of which improve patient outcomes.

FUTURE TRENDS AND APPLICATIONS

Advances in Artificial Intelligence and Image Processing

• With several trends and technologies influencing the future of AI in image processing, these fields
are continuously growing.
• Classification, object recognition, and segmentation are just a few of the jobs in image process-
ing that have been transformed by deep learning approaches, particularly convolutional neural
networks. Accuracy and efficiency will be improved by upcoming developments in architectures,
model optimization, and training methods.
• Realistic picture synthesis, style transfer, and translation are made possible by generative models
like GANs and VAEs. The ability to create and manipulate images will improve with time.

86

Advances in Artificial Intelligence for Image Processing

• For AI systems to gain public trust, explainable AI is essential. Future research should concentrate
on creating AI image processing systems that enable users to comprehend and interpret choices,
particularly in autonomous cars and medical imaging.
• By allowing collaborative training on dispersed data without raw pictures and protecting pri-
vacy while gaining from common knowledge, federated learning allays privacy concerns. The
scalability and efficiency of large-scale image processing operations will be improved by future
developments.

Emerging Applications of AI in Image Processing

AI in image processing has a wide range of uses(Alam et al., 2022; Letourneau-Guillon et al., 2020).

• Medical imaging is being revolutionised by AI-powered image processing, which helps radiologists
identify irregularities, diagnose illnesses, and create customised treatment plans. Additionally, it
helps in pathology analysis, patient outcome monitoring, and surgery planning.
• Algorithms for image processing and computer vision are essential for autonomous cars to see and
comprehend their environment. In order to provide safe and dependable driving, AI approaches
help in the recognition of objects, lanes, traffic signs, and pedestrians.
• By efficiently extracting large-scale satellite imagery, AI-based image analysis supports remote
sensing and earth observation by observing environmental changes, tracking deforestation, fore-
casting natural disasters, and evaluating ecosystem health.
• AI-powered image analysis improves video surveillance, face recognition, anomaly detection,
real-time monitoring, threat detection, and suspicious behaviour identification to increase security
in public areas, airports, and key infrastructure.
• By fusing virtual material with actual settings and enhancing picture identification, tracking, and
scene comprehension, AI approaches in AR and VR applications improve immersive experiences.

Ethical and Social Implications of AI in Image Processing

• Ethics and societal issues are raised by the use of AI in picture processing.
• Image processing by AI systems raises privacy issues, necessitating careful handling and security
of sensitive data. To solve these issues, it is imperative that laws and regulations are clear.
• Biases in training data affect image processing models, potentially leading to prejudice and bi-
assed results. Fair representation must be ensured, and biases must be reduced in training data,
algorithms, and decision-making procedures.
• AI systems must be responsible, transparent, and give users visibility into choices and processes.
Ethics standards and laws encourage ethical AI use by encouraging openness and transparency.

Industry employment may be impacted by AI-assisted image processing jobs, necessitating measures
for workforce upskilling and reskilling for new positions and possibilities. To avoid stereotypes, objec-
tionable material, and harmful representations in AI systems, image processing algorithms must respect
cultural diversity and societal standards. AI developments in image processing need to be used responsi-
bly and advantageously for society’s growth, which calls for ethical and social concerns(Boopathi, n.d.;
Koshariya, 2023; Palaniappan et al., 2023; Rahamathunnisa, 2023; Reddy et al., 2023; Sampath et al.,

87
 
Advances in Artificial Intelligence for Image Processing

2022; Selvakumar et al., 2023; Senthil et al., 2023). The development of artificial intelligence in image
processing will be influenced by new applications, moral considerations, and responsible AI use. In-
novative solutions may be created by sensibly utilising AI’s capabilities, helping to build a better future.

CONCLUSION

The basic ideas, methods, and uses of artificial intelligence in image processing are examined in this
paper. The fundamentals of digital pictures, image representation, formats, and methods for enhancing,
filtering, and restoring images are all covered. The role of artificial intelligence (AI) techniques, neural
networks, image analytics, optimization, digital watermarking, cloud computing, image augmentation,
and data pretreatment are also covered in the book chapter. Additionally, it covers data augmentation
strategies, privacy and security issues, and the advantages of cloud computing for image processing. The
work also covers emergent applications, future developments in image processing, as well as the moral
and societal ramifications of using AI for image processing. The article offers a thorough introduction
of the many facets of image processing and its applications in general.
The ability to do sophisticated analysis, automation, and decision-making has substantially changed
image processing. Future research and development will focus on advancing AI, integrating it with
cutting-edge technologies like augmented reality, virtual reality, and the internet of things (IoT), fostering
interdisciplinary collaboration between experts in image processing, computer vision, machine learning,
and domain-specific fields, ensuring ethical and responsible AI use, and deploying it in the real world.
To transform research into usable applications, these breakthroughs will need thorough validation,
exacting testing, and deployment in real-world circumstances, necessitating tight cooperation between
academics, industry, and policymakers.
Image processing is being revolutionised by artificial intelligence, which has enormous promise
across a variety of industries. Intelligent picture processing will play a significant part in determining
our future if we use AI properly and ethically. By doing so, we may open up new opportunities, handle
difficult problems, and shape this future.

ABBREVIATIONS

AI - Artificial Intelligence
CNN - Convolutional Neural Network
RNN - Recurrent Neural Network
GAN - Generative Adversarial Network
ML - Machine Learning
DL - Deep Learning
OCR - Optical Character Recognition
ROI - Region of Interest
FPS - Frames Per Second
GPU - Graphics Processing Unit
CPU - Central Processing Unit
SVM - Support Vector Machine

88

Advances in Artificial Intelligence for Image Processing

PCA - Principal Component Analysis

IoU - Intersection over Union
API - Application Programming Interface
AR - Augmented Reality
VR - Virtual Reality
NMS - Non-Maximum Suppression
SGD - Stochastic Gradient Descent
ANN - Artificial Neural Network

REFERENCES

Abduljabbar, R., Dia, H., Liyanage, S., & Bagloee, S. A. (2019). Applications of artificial intelligence
in transport: An overview. Sustainability (Basel), 11(1), 189. doi:10.3390u11010189
Alam, G., Ihsanullah, I., Naushad, M., & Sillanpää, M. (2022). Applications of artificial intelligence in
water treatment for optimization and automation of adsorption processes: Recent advances and prospects.
Chemical Engineering Journal, 427, 130011. doi:10.1016/j.cej.2021.130011
Anderson, M. R., Cafarella, M., Ros, G., & Wenisch, T. F. (2019). Physical representation-based predi-
cate optimization for a visual analytics database. 2019 IEEE 35th International Conference on Data
Engineering (ICDE), (pp. 1466–1477). IEEE.
Anitha, C., Komala, C. R., Vivekanand, C. V., Lalitha, S. D., Boopathi, S., & Revathi, R. (2023, Febru-
ary). Artificial Intelligence driven security model for Internet of Medical Things (IoMT). Proceedings
of 2023 3rd International Conference on Innovative Practices in Technology and Management, ICIPTM
2023. IEEE. 10.1109/ICIPTM57143.2023.10117713
Arnold, M., Schmucker, M., & Wolthusen, S. D. (2003). Techniques and applications of digital water-
marking and content protection. Artech House.
Babu, B. S., Kamalakannan, J., Meenatchi, N., M, S. K. S., S, K., & Boopathi, S. (2023). Economic im-
pacts and reliability evaluation of battery by adopting Electric Vehicle. IEEE Explore, 1–6. doi:10.1109/
ICPECTS56089.2022.10046786
Baduge, S. K., Thilakarathna, S., Perera, J. S., Arashpour, M., Sharafi, P., Teodosio, B., Shringi, A., &
Mendis, P. (2022). Artificial intelligence and smart vision for building and construction 4.0: Machine
and deep learning methods and applications. Automation in Construction, 141, 104440. doi:10.1016/j.
autcon.2022.104440
Bagheri, M., Akbari, A., & Mirbagheri, S. A. (2019). Advanced control of membrane fouling in filtration
systems using artificial intelligence and machine learning techniques: A critical review. Process Safety
and Environmental Protection, 123, 229–252. doi:10.1016/j.psep.2019.01.013

89
 
Advances in Artificial Intelligence for Image Processing

Berg, S., Kutra, D., Kroeger, T., Straehle, C. N., Kausler, B. X., Haubold, C., Schiegg, M., Ales, J., Beier,
T., Rudy, M., Eren, K., Cervantes, J. I., Xu, B., Beuttenmueller, F., Wolny, A., Zhang, C., Koethe, U.,
Hamprecht, F. A., & Kreshuk, A. (2019). Ilastik: Interactive machine learning for (bio) image analysis.
Nature Methods, 16(12), 1226–1232. doi:10.103841592-019-0582-9 PMID:31570887
Boopathi, S. (2019). Experimental investigation and parameter analysis of LPG refrigeration system
using Taguchi method. SN Applied Sciences, 1(8), 892. doi:10.100742452-019-0925-2
Boopathi, S. (2021). Improving of Green Sand-Mould Quality using Taguchi Technique. Journal of
Engineering Research. doi:10.36909/jer.14079
Boopathi, S. (2022a). An investigation on gas emission concentration and relative emission rate of the
near-dry wire-cut electrical discharge machining process. Environmental Science and Pollution Research
International, 29(57), 86237–86246. doi:10.100711356-021-17658-1 PMID:34837614
Boopathi, S. (2022b). Cryogenically treated and untreated stainless steel grade 317 in sustainable wire
electrical discharge machining process: A comparative study. Environmental Science and Pollution
Research International, 1–10. doi:10.100711356-022-22843-x PMID:36057706
Boopathi, S. (2022c). Experimental investigation and multi-objective optimization of cryogenic Friction-
stir-welding of AA2014 and AZ31B alloys using MOORA technique. Materials Today. Communications,
33, 104937. doi:10.1016/j.mtcomm.2022.104937
Boopathi, S. (2023a). An Investigation on Friction Stir Processing of Aluminum Alloy-Boron Carbide
Surface Composite. In Materials Horizons: From Nature to Nanomaterials (pp. 249–257). Springer.
doi:10.1007/978-981-19-7146-4_14
Boopathi, S. (2023b). Impact analysis of ceramic tile powder aggregates on self-compacting concrete.
Engineering Research Express.
Boopathi, S. (n.d.). Deep Learning Techniques Applied for Automatic Sentence Generation. IGI Global.
doi:10.4018/978-1-6684-3632-5.ch016
Boopathi, S., Arigela, S. H., Raman, R., Indhumathi, C., Kavitha, V., & Bhatt, B. C. (2023). Promi-
nent Rule Control-based Internet of Things: Poultry Farm Management System. IEEE Explore, 1–6.
doi:10.1109/ICPECTS56089.2022.10047039
Boopathi, S., Khare, R., Jaya Christiyan, K. G., Muni, T. V., & Khare, S. (2023). Additive Manufacturing
Developments in the Medical Engineering Field. In Development, Properties, and Industrial Applications
of 3D Printed Polymer Composites (pp. 86–106). IGI Global. doi:10.4018/978-1-6684-6009-2.ch006
Boopathi, S., Siva Kumar, P. K., & Meena, R. S. J., S. I., P., S. K., & Sudhakar, M. (2023). Sustainable
Developments of Modern Soil-Less Agro-Cultivation Systems. In Human Agro-Energy Optimization
for Business and Industry (pp. 69–87). IGI Global. doi:10.4018/978-1-6684-4118-3.ch004
Boopathi, S., Venkatesan, G., & Anton Savio Lewise, K. (2023). Mechanical Properties Analysis of
Kenaf–Grewia–Hair Fiber-Reinforced Composite. In Lecture Notes in Mechanical Engineering (pp.
101–110). Springer. doi:10.1007/978-981-16-9057-0_11

90

Advances in Artificial Intelligence for Image Processing

Chakravarthi, P. K., Yuvaraj, D., & Venkataramanan, V. (2022). IoT-based smart energy meter for smart
grids. ICDCS 2022 - 2022 6th International Conference on Devices, Circuits and Systems, (pp. 360–363).
IEEE. 10.1109/ICDCS54290.2022.9780714
Chen, B., Yan, Z., & Nahrstedt, K. (2022). Context-aware image compression optimization for visual
analytics offloading. Proceedings of the 13th ACM Multimedia Systems Conference, (pp. 27–38). ACM.
10.1145/3524273.3528178
Egmont-Petersen, M., de Ridder, D., & Handels, H. (2002). Image processing with neural networks—A
review. Pattern Recognition, 35(10), 2279–2301. doi:10.1016/S0031-3203(01)00178-9
Ferzli, R., & Khalife, I. (2011). Mobile cloud computing educational tool for image/video processing
algorithms. 2011 Digital Signal Processing and Signal Processing Education Meeting (DSP/SPE), (pp.
529–533). IEEE.
Gowri, N. V., Dwivedi, J. N., Krishnaveni, K., Boopathi, S., Palaniappan, M., & Medikondu, N. R.
(2023). Experimental investigation and multi-objective optimization of eco-friendly near-dry electrical
discharge machining of shape memory alloy using Cu/SiC/Gr composite electrode. Environmental Science
and Pollution Research International, 0123456789. doi:10.100711356-023-26983-6 PMID:37126160
Grupac, M. (2022). Image processing computational algorithms, sensory data mining techniques, and
predictive customer analytics in the metaverse economy. Review of Contemporary Philosophy, 21(0),
205–222. doi:10.22381/RCP21202213
Guo, W., Gong, J., Jiang, W., Liu, Y., & She, B. (2010). OpenRS-Cloud: A remote sensing image process-
ing platform based on cloud computing environment. Science China. Technological Sciences, 53(S1),
221–230. doi:10.100711431-010-3234-y
Harikaran, M., Boopathi, S., Gokulakannan, S., & Poonguzhali, M. (2023). Study on the Source of
E-Waste Management and Disposal Methods. In Sustainable Approaches and Strategies for E-Waste
Management and Utilization (pp. 39–60). IGI Global. doi:10.4018/978-1-6684-7573-7.ch003
Janardhana, K., Anushkannan, N. K., Dinakaran, K. P., Puse, R. K., & Boopathi, S. (2023). Experimental
Investigation on Microhardness, Surface Roughness, and White Layer Thickness of Dry EDM. Engineer-
ing Research Express. doi:10.1088/2631-8695/acce8f
Janardhana, K., Singh, V., Singh, S. N., Babu, T. S. R., Bano, S., & Boopathi, S. (2023). Utilization
Process for Electronic Waste in Eco-Friendly Concrete: Experimental Study. In Sustainable Approaches
and Strategies for E-Waste Management and Utilization (pp. 204–223). IGI Global.
Jeevanantham, Y. A., A, S., V, V., J, S. I., Boopathi, S., & Kumar, D. P. (2023). Implementation of Internet-of
Things (IoT) in Soil Irrigation System. IEEE Explore, 1–5. doi:10.1109/ICPECTS56089.2022.10047185
Kagadis, G. C., Kloukinas, C., Moore, K., Philbin, J., Papadimitroulas, P., Alexakos, C., Nagy, P. G.,
Visvikis, D., & Hendee, W. R. (2013). Cloud computing in medical imaging. Medical Physics, 40(7),
70901. doi:10.1118/1.4811272 PMID:23822402
Kan, A. (2017). Machine learning applications in cell image analysis. Immunology and Cell Biology,
95(6), 525–530. doi:10.1038/icb.2017.16 PMID:28294138

91
 
Advances in Artificial Intelligence for Image Processing

Kannan, E., Trabelsi, Y., Boopathi, S., & Alagesan, S. (2022). Influences of cryogenically treated work
material on near-dry wire-cut electrical discharge machining process. Surface Topography: Metrology
and Properties, 10(1), 15027. doi:10.1088/2051-672X/ac53e1
Kavitha, C., Geetha Malini, P. S., Charan Kantumuchu, V., Manoj Kumar, N., Verma, A., & Boopathi,
S. (2023). An experimental study on the hardness and wear rate of carbonitride coated stainless steel.
Materials Today: Proceedings, 74, 595–601. doi:10.1016/j.matpr.2022.09.524
Khokhar, S., Zin, A. A. B. M., Mokhtar, A. S. B., & Pesaran, M. (2015). A comprehensive overview on
signal processing and artificial intelligence techniques applications in classification of power quality
disturbances. Renewable & Sustainable Energy Reviews, 51, 1650–1663. doi:10.1016/j.rser.2015.07.068
Koshariya, A. K. (2023). AI-Enabled IoT and WSN-Integrated Smart., doi:10.4018/978-1-6684-8516-3.
ch011
Koshariya, A. K., Khatoon, S., Marathe, A. M., Suba, G. M., Baral, D., & Boopathi, S. (2023). Agri-
cultural Waste Management Systems Using Artificial Intelligence Techniques. In AI-Enabled Social
Robotics in Human Care Services (pp. 236–258). IGI Global. doi:10.4018/978-1-6684-8171-4.ch009
Kumara, V., Mohanaprakash, T. A., Fairooz, S., Jamal, K., Babu, T., & B., S. (2023). Experimental
Study on a Reliable Smart Hydroponics System. In Human Agro-Energy Optimization for Business and
Industry (pp. 27–45). IGI Global. doi:10.4018/978-1-6684-4118-3.ch002
Letourneau-Guillon, L., Camirand, D., Guilbert, F., & Forghani, R. (2020). Artificial intelligence ap-
plications for workflow, process optimization and predictive analytics. Neuroimaging Clinics of North
America, 30(4), e1–e15. doi:10.1016/j.nic.2020.08.008 PMID:33039002
Li, Z., Zhang, X., Müller, H., & Zhang, S. (2018). Large-scale retrieval for medical image analyt-
ics: A comprehensive review. Medical Image Analysis, 43, 66–84. doi:10.1016/j.media.2017.09.007
PMID:29031831
Madabhushi, A., & Lee, G. (2016). Image analysis and machine learning in digital pathology: Chal-
lenges and opportunities. Medical Image Analysis, 33, 170–175. doi:10.1016/j.media.2016.06.037
PMID:27423409
Malik, H., Srivastava, S., Sood, Y. R., & Ahmad, A. (2018). Applications of artificial intelligence tech-
niques in engineering. Sigma, 1.
Mehdy, M. M., Ng, P. Y., Shair, E. F., Saleh, N. I., & Gomes, C. (2017). Artificial neural networks in
image processing for early detection of breast cancer. Computational and Mathematical Methods in
Medicine, 2017, 2017. doi:10.1155/2017/2610628 PMID:28473865
Mikołajczyk, T., Nowicki, K., Bustillo, A., & Pimenov, D. Y. (2018). Predicting tool life in turning op-
erations using neural networks and image processing. Mechanical Systems and Signal Processing, 104,
503–513. doi:10.1016/j.ymssp.2017.11.022
Minh, T. N., Sinn, M., Lam, H. T., & Wistuba, M. (2018). Automated image data preprocessing with
deep reinforcement learning. ArXiv Preprint ArXiv:1806.05886.

92

Advances in Artificial Intelligence for Image Processing

Mohanarathinam, A., Kamalraj, S., Prasanna Venkatesan, G. K. D., Ravi, R. V., & Manikandababu, C. S.
(2020). Digital watermarking techniques for image security: A review. Journal of Ambient Intelligence
and Humanized Computing, 11(8), 3221–3229. doi:10.100712652-019-01500-1
Mohanty, A., Jothi, B., Jeyasudha, J., Ranjit, P. S., Isaac, J. S., & Boopathi, S. (2023). Additive Manu-
facturing Using Robotic Programming. In AI-Enabled Social Robotics in Human Care Services (pp.
259–282). IGI Global. doi:10.4018/978-1-6684-8171-4.ch010
Mohanty, A., Venkateswaran, N., Ranjit, P. S., Tripathi, M. A., & Boopathi, S. (2023). Innovative Strat-
egy for Profitable Automobile Industries: Working Capital Management. In Handbook of Research on
Designing Sustainable Supply Chains to Achieve a Circular Economy (pp. 412–428). IGI Global.
Mohanty, S. P. (1999). Digital watermarking: A tutorial review. CSEE. Http://Www. Csee. Usf. Edu/\~{}
Smohanty/Research/Reports/WMSurvey1999Mohanty. Pdf.
Naranjo-Torres, J., Mora, M., Hernández-García, R., Barrientos, R. J., Fredes, C., & Valenzuela, A.
(2020). A review of convolutional neural network applied to fruit image processing. Applied Sciences
(Basel, Switzerland), 10(10), 3443. doi:10.3390/app10103443
Palaniappan, M., Tirlangi, S., Mohamed, M. J. S., Moorthy, R. M. S., Valeti, S. V., & Boopathi, S. (2023).
Fused Deposition Modelling of Polylactic Acid (PLA)-Based Polymer Composites. In Development,
Properties, and Industrial Applications of 3D Printed Polymer Composites (pp. 66–85). IGI Global.
doi:10.4018/978-1-6684-6009-2.ch005
Rahamathunnisa, U. (2023). Cloud Computing Principles for Optimizing Robot Task Offloading Pro-
cesses., doi:10.4018/978-1-6684-8171-4.ch007
Rahamathunnisa, U., Sudhakar, K., Murugan, T. K., Thivaharan, S., Rajkumar, M., & Boopathi, S. (2023).
Cloud Computing Principles for Optimizing Robot Task Offloading Processes. In AI-Enabled Social
Robotics in Human Care Services (pp. 188–211). IGI Global. doi:10.4018/978-1-6684-8171-4.ch007
Ranganath, H. S., Kuntimad, G., & Johnson, J. L. (1995). Pulse coupled neural networks for image
processing. Proceedings IEEE Southeastcon’95. Visualize the Future, (pp. 37–43). IEEE.
Reddy, M. A., Reddy, B. M., Mukund, C. S., Venneti, K., Preethi, D. M. D., & Boopathi, S. (2023).
Social Health Protection During the COVID-Pandemic Using IoT. In The COVID-19 Pandemic and the
Digitalization of Diplomacy (pp. 204–235). IGI Global. doi:10.4018/978-1-7998-8394-4.ch009
Samikannu, R., Koshariya, A. K., Poornima, E., Ramesh, S., Kumar, A., & Boopathi, S. (2023). Sus-
tainable Development in Modern Aquaponics Cultivation Systems Using IoT Technologies. In Human
Agro-Energy Optimization for Business and Industry (pp. 105–127). IGI Global. doi:10.4018/978-1-
6684-4118-3.ch006
Sampath, B. C. S., & Myilsamy, S. (2022). Application of TOPSIS Optimization Technique in the Micro-
Machining Process. In Trends, Paradigms, and Advances in Mechatronics Engineering (pp. 162–187).
IGI Global. doi:10.4018/978-1-6684-5887-7.ch009

93
 
Advances in Artificial Intelligence for Image Processing

Sathish, T., Sunagar, P., Singh, V., Boopathi, S., Al-Enizi, A. M., Pandit, B., Gupta, M., & Sehgal, S. S.
(2023). Characteristics estimation of natural fibre reinforced plastic composites using deep multi-layer
perceptron (MLP) technique. Chemosphere, 337, 139346. doi:10.1016/j.chemosphere.2023.139346
PMID:37379988
Selvakumar, S., Adithe, S., Isaac, J. S., Pradhan, R., Venkatesh, V., & Sampath, B. (2023). A Study of
the Printed Circuit Board (PCB) E-Waste Recycling Process. In Sustainable Approaches and Strategies
for E-Waste Management and Utilization (pp. 159–184). IGI Global.
Senthil, T. S. R. Ohmsakthi vel, Puviyarasan, M., Babu, S. R., Surakasi, R., & Sampath, B. (2023).
Industrial Robot-Integrated Fused Deposition Modelling for the 3D Printing Process. In Development,
Properties, and Industrial Applications of 3D Printed Polymer Composites (pp. 188–210). IGI Global.
doi:10.4018/978-1-6684-6009-2.ch011
Subha, S., Inbamalar, T. M., Komala, C. R., Suresh, L. R., Boopathi, S., & Alaskar, K. (2023, February).
A Remote Health Care Monitoring system using internet of medical things (IoMT). Proceedings of 2023
3rd International Conference on Innovative Practices in Technology and Management, ICIPTM 2023.
IEEE. 10.1109/ICIPTM57143.2023.10118103
Sun, Q., Zhang, M., & Mujumdar, A. S. (2019). Recent developments of artificial intelligence in drying
of fresh food: A review. Critical Reviews in Food Science and Nutrition, 59(14), 2258–2275. doi:10.10
80/10408398.2018.1446900 PMID:29493285
Tang, S., Yuan, S., & Zhu, Y. (2020). Data preprocessing techniques in convolutional neural network
based on fault diagnosis towards rotating machinery. IEEE Access: Practical Innovations, Open Solu-
tions, 8, 149487–149496. doi:10.1109/ACCESS.2020.3012182
Tasci, E., Uluturk, C., & Ugur, A. (2021). A voting-based ensemble deep learning method focusing on
image augmentation and preprocessing variations for tuberculosis detection. Neural Computing & Ap-
plications, 33(22), 15541–15555. doi:10.100700521-021-06177-2 PMID:34121816
Vanitha, S. K. R., & Boopathi, S. (2023). Artificial Intelligence Techniques in Water Purification and
Utilization. In Human Agro-Energy Optimization for Business and Industry (pp. 202–218). IGI Global.
doi:10.4018/978-1-6684-4118-3.ch010
Vennila, T., Karuna, M. S., Srivastava, B. K., Venugopal, J., Surakasi, R., & B., S. (2023). New Strategies
in Treatment and Enzymatic Processes. In Human Agro-Energy Optimization for Business and Industry
(pp. 219–240). IGI Global. doi:10.4018/978-1-6684-4118-3.ch011
Voyatzis, G., Nikolaidis, N., & Pitas, I. (1998). Digital watermarking: An overview. 9Th European Signal
Processing Conference (EUSIPCO 1998), (pp. 1–4). IEEE.
Wang, C.-X., Di Renzo, M., Stanczak, S., Wang, S., & Larsson, E. G. (2020). Artificial intelligence
enabled wireless networking for 5G and beyond: Recent advances and future challenges. IEEE Wireless
Communications, 27(1), 16–23. doi:10.1109/MWC.001.1900292
Xu, Z., Cheng, C., & Sugumaran, V. (2020). Big data analytics of crime prevention and control based
on image processing upon cloud computing. J Surveill Secur Saf, 1, 16–33. doi:10.20517/jsss.2020.04

94
 
Advances in Artificial Intelligence for Image Processing

Xuehua, J. (2010). Digital watermarking and its application in image copyright protection. 2010 Inter-
national Conference on Intelligent Computation Technology and Automation, (vol. 2, 114–117). IEEE.
10.1109/ICICTA.2010.625
Yan, Y., & Huang, L. (2014). Large-scale image processing research cloud. Cloud Computing, 88–93.
Yuan, G., & Hao, Q. (2020). Digital watermarking secure scheme for remote sensing image protection.
China Communications, 17(4), 88–98. doi:10.23919/JCC.2020.04.009
Yupapin, P., Trabelsi, Y., Nattappan, A., & Boopathi, S. (2023). Performance Improvement of Wire-Cut
Electrical Discharge Machining Process Using Cryogenically Treated Super-Conductive State of Monel-
K500 Alloy. Iranian Journal of Science and Technology. Transaction of Mechanical Engineering, 47(1),
267–283. doi:10.100740997-022-00513-0
Zhao, H., Gallo, O., Frosio, I., & Kautz, J. (2015). Loss functions for neural networks for image process-
ing. ArXiv Preprint ArXiv:1511.08861.

95

View publication stats