Skip to main content

Manesh Kokare

Followers

13

Following

11

Public Views

Abd Rahman Ahlan

International Islamic University Malaysia

Al-Balqa Applied University

Raghu Vamsi Potukuchi

Jaypee Institute of Information Technology

Dhanalakshmi College Of Engineering

M.S.Ramaiah Institute of Technology

Kiet

Gearóid ÓLaighin

University of Galway

Svs College Of Engineering

Dr. S. VELMURUGAN M.E., Ph.D.,

Anna University

K L University

Interests

Uploads

Papers by Manesh Kokare

Composite Relevance Feedback For Image Retrieval

This paper presents content-based image retrieval (CBIR) frameworks with relevance feedback (RF) ... more This paper presents content-based image retrieval (CBIR) frameworks with relevance feedback (RF) based on combined learning of support vector machines (SVM) and AdaBoosts. The framework incorporates only most relevant images obtained from both the learning algorithm. To speed up the system, it removes irrelevant images from the database, which are returned from SVM learner. It is the key to achieve the effective retrieval performance in terms of time and accuracy. The experimental results show that this framework had significant improvement in retrieval effectiveness, which can finally improve the retrieval performance.

Automated Red Lesion Detection: An Overview

Advances in Intelligent Systems and Computing, 2020

Incidences of diabetes are increasing worldwide. An eye complication associated with uncontrolled... more Incidences of diabetes are increasing worldwide. An eye complication associated with uncontrolled diabetes is diabetic retinopathy. If not treated, diabetic retinopathy can be vision threatening. The microaneurysms and dot hemorrhages are the only prominent clinically observable symptoms of DR. Their timely detection can help ophthalmologists in treating abnormalities efficiently and limit the disease severity. So detecting red lesions in early stage has become an indispensable task today. This paper gives an overview of earlier proposed algorithms and methods. It also compares these algorithms based on their performance for supporting the researchers by providing the gist of these algorithms. The standard retinal image databases are also compared and discussed.

Automatic Retinal Image Analysis for the Detection of Diabetic Retinopathy

Biomedical Signal and Image Processing in Patient Care, 2018

Diabetic Retinopathy, a condition in the person affected by diabetes, is most common cause of bli... more Diabetic Retinopathy, a condition in the person affected by diabetes, is most common cause of blindness in the world. Recent research has given a better understanding of requirement in clinical eye care practice to identify better and cheaper ways of identification, management, diagnosis and treatment of retinal disease. The importance of diabetic retinopathy screening programs and difficulty in achieving reliable early diagnosis of diabetic retinopathy at a reasonable cost needs attention to develop computer-aided diagnosis tool. Computer aided disease diagnosis in retinal image analysis could ease mass screening of population with diabetes mellitus and help clinicians in utilizing their time more efficiently. The recent technological advances in computing power, communication systems, and machine learning techniques provide opportunities to the biomedical engineers and computer scientists to meet the requirements of clinical practice. With proper self-care, management, and medical...

Automatic Field of View Extraction with Variable Enhancement of Color Fundus Images

2017 14th IEEE India Council International Conference (INDICON), 2017

Most retinal images suffers from large size with inadequate contrast and brightness. Automated fi... more Most retinal images suffers from large size with inadequate contrast and brightness. Automated field of view (FOV) extraction and image enhancement significantly increases the performance of image analysis for diabetic retinopathy (DR) detection. This paper presents a simple and automated FOV extraction technique using only one horizontal and vertical scan lines. The FOV image further enhanced using the guided filter with variable enhancement factor. The proposed method has been applied on five publicly available databases. The accuracy of FOV extraction is obtained 100% for Diaretdb0 and ONHSD databases, 99.91% for messidor, 99.32% for E-optha and 98.87% for Diaretdb1 database. The experimental results shows that, proposed method gives the good enhancement on five different databases compared with the existing state-of-the-art methods. Overall, the proposed method can be used as a effective preprocessing tool for accurate detection of lesions in retinal image analysis.

Emantic I Mage R Etrieval Using Relevance Feedback

This paper presents optimized interactive content-based image retrieval framework based on AdaBoo... more This paper presents optimized interactive content-based image retrieval framework based on AdaBoost learning method. As we know relevance feedback (RF) is online process, so we have optimized the learning process by considering the most positive image selection on each feedback iteration. To learn the system we have used AdaBoost. The main significances of our system are to address the small training sample and to reduce retrieval time. Experiments are conducted on 1000 semantic colour images from Corel database to demonstrate the effectiveness of the proposed framework. These experiments employed large image database and combined RCWFs and DT-CWT texture descriptors to represent content of the images.

WheatNet-Lite: A Novel Light Weight Network for Wheat Head Detection

2021 IEEE/CVF International Conference on Computer Vision Workshops (ICCVW), 2021

Recently, the potential for wheat head detection has been significantly enhanced using deep learn... more Recently, the potential for wheat head detection has been significantly enhanced using deep learning techniques. However, the significant challenges are variation in growth stages of wheat heads, canopy, genotype, and wheat head orientation. Furthermore, the wheat head detection task gets even more complex due to the overlapping density of wheat heads and the blur image due to the wind. For real-time wheat head detection, designing lightweight deep learning models for edge devices is also challenging. This paper proposes a lightweight WheatNet-Lite architecture to enhance the efficiency and accuracy of wheat head detection. The proposed method utilizes Mixed Depthwise Conv (MDWConv) with an inverted residual bottleneck in the backbone. Additionally, the Modified Spatial Pyramidal Polling (MSPP) effectively extracts the multi-scale features. The final wheat head bounding box prediction is achieved using WheatNet-lite Neck by utilizing Depthwise Convolution (DWConv) with a Feature Pyramid structure. It reduces 54.2 M network parameters in comparison to YOLOV3. The proposed approach outperforms the existing state-of-the-art methods with mean average precision (mAP) of 91.32 [email protected] and 86.10 [email protected] on GWHD and SPIKE datasets, respectively, with only 8.2 M parameters. Also, the new ACID dataset is proposed with bounding box annotation with 76.32 [email protected]. The experimental results are demonstrated on three different datasets viz. Global Wheat Head Detection (GWHD) , SPIKE dataset, and Annotated Crop Image Dataset (ACID) showing a significant improvement in the wheat head detection with speed and accuracy.

Writer based Handwritten Document Image Retrieval

this paper a method is proposed for retrieval of handwritten document images based on the writer&... more this paper a method is proposed for retrieval of handwritten document images based on the writer's handwriting using texture features of input handwritten document image block. Typically it can be observed that the patterns of any handwritten text blocks encompass spatial texture primitives. The conventional two-dimensional (2-D) discrete wavelet transforms (DWTs) and Correlation of GLCM is used to extract spatial features. Handwritten documents are collected from 100 writers each in English, Kannada and Hindi scripts. These handwritten documents are segmented into image blocks and 2000 image blocks of each script writers are used separately for validation of the proposed method. The similarity measures viz., Euclidean and City block distances are used and achieved Top-1 retrieval rates as 100% for each of the Kannada, English and Hindi writers' document image blocks.

Automatic Segmentation of Optic Disc by Gradient Minimization Based Approach

2018 International Conference on Intelligent and Advanced System (ICIAS), 2018

Diabetic Retinopathy (DR) and Glaucoma are the eye diseases that cause progressive damage to pati... more Diabetic Retinopathy (DR) and Glaucoma are the eye diseases that cause progressive damage to patient's vision over time. Optic disc detection and segmentation is of significant importance for automating the screening of both diseases. The method proposed for the precise and fast segmentation of the optic disc mainly based obtaining optic disc center location using intensity profile analysis and further segmentation of chosen region of interest using $L_{0}$ gradient minimization based technique. The optimum threshold was chosen for OD segmentation using Otsu's method. Performance of the technique is evaluated using Dice and Jaccard metrics that are obtained as 80.23% and 71.17% 78.46% and 82.63% respectively on IDRiD and the MESSIDOR dataset. The proposed method is found to be very fast and efficient.

IDRiD: Diabetic Retinopathy – Segmentation and Grading Challenge

Medical Image Analysis, 2019

challenge. This paper outlines the challenge, its organization, the dataset used, evaluation meth... more challenge. This paper outlines the challenge, its organization, the dataset used, evaluation methods and results of top performing participating solutions. We observe that the top performing approaches utilize a blend of clinical information, data augmentation, and the ensemble of models. These findings have the potential to enable new developments in retinal image analysis and image-based DR screening in particular.

A heuristic optimisation approach for semantic-based image retrieval using relevance feedback

International Journal of Applied Pattern Recognition, 2016

Relevance feedback has been considered as the most vivid approach for reduction of semantic gap i... more Relevance feedback has been considered as the most vivid approach for reduction of semantic gap in content-based image retrieval systems. However, existing relevance feedback techniques require more number of feedback iterations to fulfil user's requirement. To address the problem of convergence speed in relevance feedback, this paper proposes a novel approach using bacterial foraging optimisation algorithm. The proposed approach combines the query point movement and feature relevance weighting techniques in relevance feedback. The feature weights in feature relevance weighting technique are obtained using a heuristic approach based on bacterial foraging optimisation algorithm. The proposed system is tested on two different image databases. The experimental results confirm the high accuracy and effectiveness of the proposed system as compared to other content-based image retrieval systems available in the literature.

Human age classification using appearance features and artificial neural network

International Journal of Biometrics, 2016

This paper presents a novel method for human age classification via face images by a computer. Th... more This paper presents a novel method for human age classification via face images by a computer. The proposed method classifies the human face images into four age groups: child, young, adult and senior adult by using appearance features as ageing features and artificial neural network (ANN) as age classifier. The appearance features consist of both shape and textural features. Only two geometric ratios in combination with newly introduced rotation, scale and translation invariant efficient feature face angle are used as shape features. Local binary pattern histogram (LBPH) of regions of interest in face images are used as textural features. The ANN is designed by using two layer feedforward backpropagation neural networks. The performance of proposed age classification system is evaluated on face images from FG-NET ageing database and achieved greatly improved accuracy of 91.09% and 88.18% for male and female, respectively.

Human age classification using facial skin aging features and artificial neural network

Cognitive Systems Research, 2016

In this paper a novel method based on facial skin aging features and Artificial Neural Network (A... more In this paper a novel method based on facial skin aging features and Artificial Neural Network (ANN) is proposed to classify the human face images into four age groups. The facial skin aging features are extracted by using Local Gabor Binary Pattern Histogram (LGBPH) and wrinkle analysis. The ANN classifier is designed by using two layer feedforward backpropagation neural networks. The proposed age classification framework is trained and tested with face images from PAL face database and shown considerable improvement in the age classification accuracy up to 94.17% and 93.75% for male and female respectively.

Iris Recognition Using Discrete Wavelet Transform

2009 International Conference on Digital Image Processing, 2009

We propose a novel approach for improved iris recognition system with reduced false acceptance ra... more We propose a novel approach for improved iris recognition system with reduced false acceptance rate (FAR) and false rejection rate (FRR). The technique developed here uses all the frequency resolution planes of discrete wavelet transform (DWT). These frequency planes provide abundant texture information present in an iris at different resolutions. The accuracy is improved up to 98.98%. With proposed method

Rotation-invariant fingerprint matching using radon and DCT

Sādhanā, 2017

A new set of promising rotation-invariant features based on radon and discrete cosine transform (... more A new set of promising rotation-invariant features based on radon and discrete cosine transform (DCT) is proposed for fingerprint matching. The radon and DCT of a tiny area in the region of core point of fingerprint image is computed. In the proposed method only 34% DCT coefficients are used for feature extraction. Competency of this approach is tested on standard databases, namely FVC2002 and FVC2004. This approach provides 70% genuine acceptance rate (GAR) at *0% false acceptance rate (FAR) and 95% GAR at 10% FAR on rotated and non-rotated databases, respectively. Experimental results prove that the proposed feature extraction approach is rotation invariant.

Content-Based Image Retrieval with Relevance Feedback Using Adaptive Processing of Tree-Structure Image Representation

International Journal of Image and Graphics, 2003

Content-based image retrieval has become an essential technique in multimedia data management. Ho... more Content-based image retrieval has become an essential technique in multimedia data management. However, due to the difficulties and complications involved in the various image processing tasks, a robust semantic representation of image content is still very difficult (if not impossible) to achieve. In this paper, we propose a novel content-based image retrieval approach with relevance feedback using adaptive processing of tree-structure image representation. In our approach, each image is first represented with a quad-tree, which is segmentation free. Then a neural network model with the Back-Propagation Through Structure (BPTS) learning algorithm is employed to learn the tree-structure representation of the image content. This approach that integrates image representation and similarity measure in a single framework is applied to the relevance feedback of the content-based image retrieval. In our approach, an initial ranking of the database images is first carried out based on the ...

Automated Diabetic Macular Edema (DME) Analysis using Fine Tuning with Inception-Resnet-v2 on OCT Images

2018 IEEE-EMBS Conference on Biomedical Engineering and Sciences (IECBES), 2018

Accurate detection of diabetic macular edema (DME) is an important task in optical coherence tomo... more Accurate detection of diabetic macular edema (DME) is an important task in optical coherence tomography (OCT) images of the eye. A relatively simple and practical approach is proposed in this paper. A pre-trained convolutional neural network (CNN) is fine tuned for a classification of DME versus normal cases. The fine-tuned Inception-Resnet-v2 CNN model can effectively identify pathologies in comparison to classical learning. Experiments were carried out on the publicly available data set of the Singapore Eye Research Institute (SERI). The developed model was also compared to other fine tuned models, such as Resnet-50 and Inception-v3. The proposed method achieved 100% classification accuracy with the Inception-Resnet-v2 model using a leave-one-out crossvalidation strategy. For robustness, the model trained on the SERI dataset was tested on another dataset provided by the Chinese University HongKong (CUHK), also with 100% accuracy. The proposed method is a potentially impactful tool for accurately detecting DME vs. normal cases.

Fingerprint Matching Based on Dimension Reduced DCT Feature Vectors

Journal of Information Processing Systems

In this work a Discrete Cosine Transform (DCT)-based feature dimensionality reduced approach for ... more In this work a Discrete Cosine Transform (DCT)-based feature dimensionality reduced approach for fingerprint matching is proposed. The DCT is applied on a small region around the core point of fingerprint image. The performance of our proposed method is evaluated on a small database of Bologna University and two large databases of FVC2000. A dimensionally reduced feature vector is formed using only approximately 19%, 7%, and 6% DCT coefficients for the three databases from Bologna University and FVC2000, respectively. We compared the results of our proposed method with the discrete wavelet transform (DWT) method, the rotated wavelet filters (RWFs) method, and a combination of DWT+RWF and DWT+(HL+LH) subbands of RWF. The proposed method reduces the false acceptance rate from approximately 18% to 4% on DB1 (Database of Bologna University), approximately 29% to 16% on DB2 (FVC2000), and approximately 26% to 17% on DB3 (FVC2000) over the DWT based feature extraction method.

NENet: Nested EfficientNet and Adversarial Learning for Joint Optic Disc and Cup Segmentation

Medical Image Analysis

Retinal Fundus Multi-Disease Image Dataset (RFMiD): A Dataset for Multi-Disease Detection Research

Data

The world faces difficulties in terms of eye care, including treatment, quality of prevention, vi... more The world faces difficulties in terms of eye care, including treatment, quality of prevention, vision rehabilitation services, and scarcity of trained eye care experts. Early detection and diagnosis of ocular pathologies would enable forestall of visual impairment. One challenge that limits the adoption of computer-aided diagnosis tool by ophthalmologists is the number of sight-threatening rare pathologies, such as central retinal artery occlusion or anterior ischemic optic neuropathy, and others are usually ignored. In the past two decades, many publicly available datasets of color fundus images have been collected with a primary focus on diabetic retinopathy, glaucoma, age-related macular degeneration and few other frequent pathologies. To enable development of methods for automatic ocular disease classification of frequent diseases along with the rare pathologies, we have created a new Retinal Fundus Multi-disease Image Dataset (RFMiD). It consists of 3200 fundus images captured ...

Random forest-based active learning for content-based image retrieval

International Journal of Intelligent Information and Database Systems

Composite Relevance Feedback For Image Retrieval

This paper presents content-based image retrieval (CBIR) frameworks with relevance feedback (RF) ... more This paper presents content-based image retrieval (CBIR) frameworks with relevance feedback (RF) based on combined learning of support vector machines (SVM) and AdaBoosts. The framework incorporates only most relevant images obtained from both the learning algorithm. To speed up the system, it removes irrelevant images from the database, which are returned from SVM learner. It is the key to achieve the effective retrieval performance in terms of time and accuracy. The experimental results show that this framework had significant improvement in retrieval effectiveness, which can finally improve the retrieval performance.

Automated Red Lesion Detection: An Overview

Advances in Intelligent Systems and Computing, 2020

Incidences of diabetes are increasing worldwide. An eye complication associated with uncontrolled... more Incidences of diabetes are increasing worldwide. An eye complication associated with uncontrolled diabetes is diabetic retinopathy. If not treated, diabetic retinopathy can be vision threatening. The microaneurysms and dot hemorrhages are the only prominent clinically observable symptoms of DR. Their timely detection can help ophthalmologists in treating abnormalities efficiently and limit the disease severity. So detecting red lesions in early stage has become an indispensable task today. This paper gives an overview of earlier proposed algorithms and methods. It also compares these algorithms based on their performance for supporting the researchers by providing the gist of these algorithms. The standard retinal image databases are also compared and discussed.

Automatic Retinal Image Analysis for the Detection of Diabetic Retinopathy

Biomedical Signal and Image Processing in Patient Care, 2018

Diabetic Retinopathy, a condition in the person affected by diabetes, is most common cause of bli... more Diabetic Retinopathy, a condition in the person affected by diabetes, is most common cause of blindness in the world. Recent research has given a better understanding of requirement in clinical eye care practice to identify better and cheaper ways of identification, management, diagnosis and treatment of retinal disease. The importance of diabetic retinopathy screening programs and difficulty in achieving reliable early diagnosis of diabetic retinopathy at a reasonable cost needs attention to develop computer-aided diagnosis tool. Computer aided disease diagnosis in retinal image analysis could ease mass screening of population with diabetes mellitus and help clinicians in utilizing their time more efficiently. The recent technological advances in computing power, communication systems, and machine learning techniques provide opportunities to the biomedical engineers and computer scientists to meet the requirements of clinical practice. With proper self-care, management, and medical...

Automatic Field of View Extraction with Variable Enhancement of Color Fundus Images

2017 14th IEEE India Council International Conference (INDICON), 2017

Most retinal images suffers from large size with inadequate contrast and brightness. Automated fi... more Most retinal images suffers from large size with inadequate contrast and brightness. Automated field of view (FOV) extraction and image enhancement significantly increases the performance of image analysis for diabetic retinopathy (DR) detection. This paper presents a simple and automated FOV extraction technique using only one horizontal and vertical scan lines. The FOV image further enhanced using the guided filter with variable enhancement factor. The proposed method has been applied on five publicly available databases. The accuracy of FOV extraction is obtained 100% for Diaretdb0 and ONHSD databases, 99.91% for messidor, 99.32% for E-optha and 98.87% for Diaretdb1 database. The experimental results shows that, proposed method gives the good enhancement on five different databases compared with the existing state-of-the-art methods. Overall, the proposed method can be used as a effective preprocessing tool for accurate detection of lesions in retinal image analysis.

Emantic I Mage R Etrieval Using Relevance Feedback

This paper presents optimized interactive content-based image retrieval framework based on AdaBoo... more This paper presents optimized interactive content-based image retrieval framework based on AdaBoost learning method. As we know relevance feedback (RF) is online process, so we have optimized the learning process by considering the most positive image selection on each feedback iteration. To learn the system we have used AdaBoost. The main significances of our system are to address the small training sample and to reduce retrieval time. Experiments are conducted on 1000 semantic colour images from Corel database to demonstrate the effectiveness of the proposed framework. These experiments employed large image database and combined RCWFs and DT-CWT texture descriptors to represent content of the images.

WheatNet-Lite: A Novel Light Weight Network for Wheat Head Detection

2021 IEEE/CVF International Conference on Computer Vision Workshops (ICCVW), 2021

Recently, the potential for wheat head detection has been significantly enhanced using deep learn... more Recently, the potential for wheat head detection has been significantly enhanced using deep learning techniques. However, the significant challenges are variation in growth stages of wheat heads, canopy, genotype, and wheat head orientation. Furthermore, the wheat head detection task gets even more complex due to the overlapping density of wheat heads and the blur image due to the wind. For real-time wheat head detection, designing lightweight deep learning models for edge devices is also challenging. This paper proposes a lightweight WheatNet-Lite architecture to enhance the efficiency and accuracy of wheat head detection. The proposed method utilizes Mixed Depthwise Conv (MDWConv) with an inverted residual bottleneck in the backbone. Additionally, the Modified Spatial Pyramidal Polling (MSPP) effectively extracts the multi-scale features. The final wheat head bounding box prediction is achieved using WheatNet-lite Neck by utilizing Depthwise Convolution (DWConv) with a Feature Pyramid structure. It reduces 54.2 M network parameters in comparison to YOLOV3. The proposed approach outperforms the existing state-of-the-art methods with mean average precision (mAP) of 91.32 [email protected] and 86.10 [email protected] on GWHD and SPIKE datasets, respectively, with only 8.2 M parameters. Also, the new ACID dataset is proposed with bounding box annotation with 76.32 [email protected]. The experimental results are demonstrated on three different datasets viz. Global Wheat Head Detection (GWHD) , SPIKE dataset, and Annotated Crop Image Dataset (ACID) showing a significant improvement in the wheat head detection with speed and accuracy.

Writer based Handwritten Document Image Retrieval

this paper a method is proposed for retrieval of handwritten document images based on the writer&... more this paper a method is proposed for retrieval of handwritten document images based on the writer's handwriting using texture features of input handwritten document image block. Typically it can be observed that the patterns of any handwritten text blocks encompass spatial texture primitives. The conventional two-dimensional (2-D) discrete wavelet transforms (DWTs) and Correlation of GLCM is used to extract spatial features. Handwritten documents are collected from 100 writers each in English, Kannada and Hindi scripts. These handwritten documents are segmented into image blocks and 2000 image blocks of each script writers are used separately for validation of the proposed method. The similarity measures viz., Euclidean and City block distances are used and achieved Top-1 retrieval rates as 100% for each of the Kannada, English and Hindi writers' document image blocks.

Automatic Segmentation of Optic Disc by Gradient Minimization Based Approach

2018 International Conference on Intelligent and Advanced System (ICIAS), 2018

Diabetic Retinopathy (DR) and Glaucoma are the eye diseases that cause progressive damage to pati... more Diabetic Retinopathy (DR) and Glaucoma are the eye diseases that cause progressive damage to patient's vision over time. Optic disc detection and segmentation is of significant importance for automating the screening of both diseases. The method proposed for the precise and fast segmentation of the optic disc mainly based obtaining optic disc center location using intensity profile analysis and further segmentation of chosen region of interest using $L_{0}$ gradient minimization based technique. The optimum threshold was chosen for OD segmentation using Otsu's method. Performance of the technique is evaluated using Dice and Jaccard metrics that are obtained as 80.23% and 71.17% 78.46% and 82.63% respectively on IDRiD and the MESSIDOR dataset. The proposed method is found to be very fast and efficient.

IDRiD: Diabetic Retinopathy – Segmentation and Grading Challenge

Medical Image Analysis, 2019

challenge. This paper outlines the challenge, its organization, the dataset used, evaluation meth... more challenge. This paper outlines the challenge, its organization, the dataset used, evaluation methods and results of top performing participating solutions. We observe that the top performing approaches utilize a blend of clinical information, data augmentation, and the ensemble of models. These findings have the potential to enable new developments in retinal image analysis and image-based DR screening in particular.

A heuristic optimisation approach for semantic-based image retrieval using relevance feedback

International Journal of Applied Pattern Recognition, 2016

Relevance feedback has been considered as the most vivid approach for reduction of semantic gap i... more Relevance feedback has been considered as the most vivid approach for reduction of semantic gap in content-based image retrieval systems. However, existing relevance feedback techniques require more number of feedback iterations to fulfil user's requirement. To address the problem of convergence speed in relevance feedback, this paper proposes a novel approach using bacterial foraging optimisation algorithm. The proposed approach combines the query point movement and feature relevance weighting techniques in relevance feedback. The feature weights in feature relevance weighting technique are obtained using a heuristic approach based on bacterial foraging optimisation algorithm. The proposed system is tested on two different image databases. The experimental results confirm the high accuracy and effectiveness of the proposed system as compared to other content-based image retrieval systems available in the literature.

Human age classification using appearance features and artificial neural network

International Journal of Biometrics, 2016

This paper presents a novel method for human age classification via face images by a computer. Th... more This paper presents a novel method for human age classification via face images by a computer. The proposed method classifies the human face images into four age groups: child, young, adult and senior adult by using appearance features as ageing features and artificial neural network (ANN) as age classifier. The appearance features consist of both shape and textural features. Only two geometric ratios in combination with newly introduced rotation, scale and translation invariant efficient feature face angle are used as shape features. Local binary pattern histogram (LBPH) of regions of interest in face images are used as textural features. The ANN is designed by using two layer feedforward backpropagation neural networks. The performance of proposed age classification system is evaluated on face images from FG-NET ageing database and achieved greatly improved accuracy of 91.09% and 88.18% for male and female, respectively.

Human age classification using facial skin aging features and artificial neural network

Cognitive Systems Research, 2016

In this paper a novel method based on facial skin aging features and Artificial Neural Network (A... more In this paper a novel method based on facial skin aging features and Artificial Neural Network (ANN) is proposed to classify the human face images into four age groups. The facial skin aging features are extracted by using Local Gabor Binary Pattern Histogram (LGBPH) and wrinkle analysis. The ANN classifier is designed by using two layer feedforward backpropagation neural networks. The proposed age classification framework is trained and tested with face images from PAL face database and shown considerable improvement in the age classification accuracy up to 94.17% and 93.75% for male and female respectively.

Iris Recognition Using Discrete Wavelet Transform

2009 International Conference on Digital Image Processing, 2009

We propose a novel approach for improved iris recognition system with reduced false acceptance ra... more We propose a novel approach for improved iris recognition system with reduced false acceptance rate (FAR) and false rejection rate (FRR). The technique developed here uses all the frequency resolution planes of discrete wavelet transform (DWT). These frequency planes provide abundant texture information present in an iris at different resolutions. The accuracy is improved up to 98.98%. With proposed method

Rotation-invariant fingerprint matching using radon and DCT

Sādhanā, 2017

A new set of promising rotation-invariant features based on radon and discrete cosine transform (... more A new set of promising rotation-invariant features based on radon and discrete cosine transform (DCT) is proposed for fingerprint matching. The radon and DCT of a tiny area in the region of core point of fingerprint image is computed. In the proposed method only 34% DCT coefficients are used for feature extraction. Competency of this approach is tested on standard databases, namely FVC2002 and FVC2004. This approach provides 70% genuine acceptance rate (GAR) at *0% false acceptance rate (FAR) and 95% GAR at 10% FAR on rotated and non-rotated databases, respectively. Experimental results prove that the proposed feature extraction approach is rotation invariant.

Content-Based Image Retrieval with Relevance Feedback Using Adaptive Processing of Tree-Structure Image Representation

International Journal of Image and Graphics, 2003

Content-based image retrieval has become an essential technique in multimedia data management. Ho... more Content-based image retrieval has become an essential technique in multimedia data management. However, due to the difficulties and complications involved in the various image processing tasks, a robust semantic representation of image content is still very difficult (if not impossible) to achieve. In this paper, we propose a novel content-based image retrieval approach with relevance feedback using adaptive processing of tree-structure image representation. In our approach, each image is first represented with a quad-tree, which is segmentation free. Then a neural network model with the Back-Propagation Through Structure (BPTS) learning algorithm is employed to learn the tree-structure representation of the image content. This approach that integrates image representation and similarity measure in a single framework is applied to the relevance feedback of the content-based image retrieval. In our approach, an initial ranking of the database images is first carried out based on the ...

Automated Diabetic Macular Edema (DME) Analysis using Fine Tuning with Inception-Resnet-v2 on OCT Images

2018 IEEE-EMBS Conference on Biomedical Engineering and Sciences (IECBES), 2018

Accurate detection of diabetic macular edema (DME) is an important task in optical coherence tomo... more Accurate detection of diabetic macular edema (DME) is an important task in optical coherence tomography (OCT) images of the eye. A relatively simple and practical approach is proposed in this paper. A pre-trained convolutional neural network (CNN) is fine tuned for a classification of DME versus normal cases. The fine-tuned Inception-Resnet-v2 CNN model can effectively identify pathologies in comparison to classical learning. Experiments were carried out on the publicly available data set of the Singapore Eye Research Institute (SERI). The developed model was also compared to other fine tuned models, such as Resnet-50 and Inception-v3. The proposed method achieved 100% classification accuracy with the Inception-Resnet-v2 model using a leave-one-out crossvalidation strategy. For robustness, the model trained on the SERI dataset was tested on another dataset provided by the Chinese University HongKong (CUHK), also with 100% accuracy. The proposed method is a potentially impactful tool for accurately detecting DME vs. normal cases.

Fingerprint Matching Based on Dimension Reduced DCT Feature Vectors

Journal of Information Processing Systems

In this work a Discrete Cosine Transform (DCT)-based feature dimensionality reduced approach for ... more In this work a Discrete Cosine Transform (DCT)-based feature dimensionality reduced approach for fingerprint matching is proposed. The DCT is applied on a small region around the core point of fingerprint image. The performance of our proposed method is evaluated on a small database of Bologna University and two large databases of FVC2000. A dimensionally reduced feature vector is formed using only approximately 19%, 7%, and 6% DCT coefficients for the three databases from Bologna University and FVC2000, respectively. We compared the results of our proposed method with the discrete wavelet transform (DWT) method, the rotated wavelet filters (RWFs) method, and a combination of DWT+RWF and DWT+(HL+LH) subbands of RWF. The proposed method reduces the false acceptance rate from approximately 18% to 4% on DB1 (Database of Bologna University), approximately 29% to 16% on DB2 (FVC2000), and approximately 26% to 17% on DB3 (FVC2000) over the DWT based feature extraction method.

NENet: Nested EfficientNet and Adversarial Learning for Joint Optic Disc and Cup Segmentation

Medical Image Analysis

Retinal Fundus Multi-Disease Image Dataset (RFMiD): A Dataset for Multi-Disease Detection Research

Data

The world faces difficulties in terms of eye care, including treatment, quality of prevention, vi... more The world faces difficulties in terms of eye care, including treatment, quality of prevention, vision rehabilitation services, and scarcity of trained eye care experts. Early detection and diagnosis of ocular pathologies would enable forestall of visual impairment. One challenge that limits the adoption of computer-aided diagnosis tool by ophthalmologists is the number of sight-threatening rare pathologies, such as central retinal artery occlusion or anterior ischemic optic neuropathy, and others are usually ignored. In the past two decades, many publicly available datasets of color fundus images have been collected with a primary focus on diabetic retinopathy, glaucoma, age-related macular degeneration and few other frequent pathologies. To enable development of methods for automatic ocular disease classification of frequent diseases along with the rare pathologies, we have created a new Retinal Fundus Multi-disease Image Dataset (RFMiD). It consists of 3200 fundus images captured ...

Random forest-based active learning for content-based image retrieval

International Journal of Intelligent Information and Database Systems