Papers by Md Humaion Kabir Mehedi
6th International Conference on Electrical Engineering and Information & Communication Technology (ICEEICT) 02-04 May 2024, Military Institute of Science and Technology (MIST), Dhaka-1216, Bangladesh, 2024
As the digital landscape expands, the rise of online hate speech presents a pressing challenge, n... more As the digital landscape expands, the rise of online hate speech presents a pressing challenge, necessitating sophisticated tools for effective detection and mitigation. This project focuses on the intricate linguistic landscape of Banglish a hybrid language amalgamating Bengali and English striving to develop robust models tailored to its unique characteristics. The dataset, comprising 5000 Banglish comments categorized into various hate speech types, serves as the foundation for model exploration. Our approach spans a wide variety of models, including traditional machine learning (SVM, Logistic Regression,random forest), advanced deep learning architectures and innovative hybrid models (CNN+BiLSTM). Approaches for feature extraction such word embedding, TF-IDF, and Bag-of-Words and sentiment analysis scores are adapted to the nuances of Banglish. Ethical considerations guide our development, addressing algorithmic bias and user rights. The experimental results provide a nuanced understanding of model performance, including accuracy (90%), precision, recall, and F1 score. Insights derived from these analyses contribute to the ongoing refinement of hate speech detection methodologies, advancing the field of computational linguistics and ethical artificial intelligence.
6th International Conference on Electrical Engineering and Information & Communication Technology (ICEEICT) 02-04 May 2024, Military Institute of Science and Technology (MIST), Dhaka-1216, Bangladesh, 2024
Online issues including hate speech, abusive communications, and harassment have been exacerbated... more Online issues including hate speech, abusive communications, and harassment have been exacerbated by the rising number of Internet users. People in Bangladesh often face online harassment and threats expressed in Bengali on various social media platforms. Also, there has not been nearly enough investigation into the possibility of Offensive language in Bengali literature. Although finding realistic ways to reduce hate speech in Bengali texts is urgently needed, there is a notable lack of study in the area of Bengali abusive speech detection, despite the widespread detrimental impacts of abusive text on people's wellbeing. The results of this research provide a method for spotting bad hateful comments in Bengali online profiles. This research provides a methodology to identify potentially manipulative hate speech in Bengali social media postings. The BERT architecture is used to gather characteristics of Bengali texts. The next step in hate speech classification is to use a Convolutional Neural Network (CNN) model including a softMax activation function. We propose a new model, BERT-CNN, that combines both models. On the Bengali Hate Speech from Social Platforms (BD-SHS) dataset, the BERT-CNN model outperformed most baseline architectures, with accuracy, precision, recall, and F1-scores of 95.67%, 93.55%,92.67%, and 94.44%, respectively. According to our research, the method we suggested for spotting hate speech in Bengali writings posted on social networking sites works well, which can lessen online hate comments and foster a more civilized online community.
2022 IEEE International Conference on Artificial Intelligence in Engineering and Technology (IICAIET)
ICCIT, 2023
Emotions are a critical factor in intrapersonal communication and significantly influence how we ... more Emotions are a critical factor in intrapersonal communication and significantly influence how we convey our intentions and feelings. Recognizing emotion through speech not only enhances our understanding of interpersonal dynamics but also holds immense potential across diverse sectors such as healthcare, human-machine interaction, automated customer service and more. However, the majority of the existing speech recognition systems are highly centralized, raising concerns over potential data leakage. To address this issue, we have introduced a privacy preserving system to recognize emotion from audio data using federated learning. Our approach leveraged the distributed model training and aggregation strategy, ensuring data privacy while eliminating the need for data sharing to a centralized system. Moreover, we explored the potential of CNN and LSTM models, both as a distributed and centralized formats, with MFCC as features in the federated learning setting. Experimental evaluation of our approach on the IEMOCAP and CREMA-D datasets achieved a maximum accuracy of 68.65% and 68.82%, surpassing the existing federated learning techniques and rivaling centralized benchmarks.
ICCIT, 2023
Visual Question Answering (VQA) is a field where computer vision and natural language processing ... more Visual Question Answering (VQA) is a field where computer vision and natural language processing intersect to develop systems capable of comprehending visual information and answering natural language questions. In visual question answering , algorithms interpret real-world images in response to questions expressed in human language. Our paper presents an extensive experimental study on Visual Question Answering (VQA) using a diverse set of multimodal transformers. The VQA task requires systems to comprehend both visual content and natural language questions. To address this challenge, we explore the performance of various pre-trained transformer architectures for encoding questions, including BERT, RoBERTa, and ALBERT, as well as image transformers, such as ViT, DeiT, and BEiT, for encoding images. Multimodal transformers' smooth fusion of visual and text data promotes cross-modal understanding and strengthens reasoning skills. On benchmark datasets like the Visual Question Answering (VQA) v2.0 dataset, we rigorously test and fine-tune these models to assess their effectiveness and compare their performance to more conventional VQA methods. The results show that multimodal transformers significantly outperform traditional techniques in terms of performance. Additionally, the models' attention maps give users insights into how they make decisions, improving interpretability and comprehension. Because of their adaptability, the tested transformer topologies have the potential to be used in a wide range of VQA applications, such as robotics, healthcare, and assistive technology. This study demonstrates the effectiveness and promise of multimodal transformers as a method for improving the effectiveness of visual question-answering systems.
ICCIT, 2023
Reviews can significantly impact a company's reputation in the market, potentially influencing it... more Reviews can significantly impact a company's reputation in the market, potentially influencing its overall business outcomes, either positively or negatively. This is especially crucial for companies that operate primarily through e-commerce platforms. Hence, it is vital for companies to pay close attention to customer reviews. Sentiment Analysis, often referred to as "opinion mining," is a significant procedure in Natural Language Processing (NLP) which serves the purpose of ascertaining the emotional tone of a provided text and categorizing it into positive, negative, or neutral perspectives. In this paper, sentiment analysis methodology is presented for classifying Amazon reviews which utilizes a large dataset of reviews and employs Multinomial Naïve Bayesian (MNB), Support Vector Machine (SVM), Maximum Entropy (ME), and Logistic Regression as the primary classifiers by the authors. With the aid of machine learning, we employed a supervised learning approach to an extensive Amazon dataset in order to categorize it based on sentiment polarity, achieving a high level of accuracy for the results. Here, we utilized the Kaggle dataset that includes a substantial volume of reviews and associated metadata which comprises customer reviews and ratings on Amazon products.
ICCIT, 2023
Handwritten signature verification is a crucial task with applications spanning authentication, f... more Handwritten signature verification is a crucial task with applications spanning authentication, financial transactions, and legal documents. In scenarios where only a single reference signature is available, the challenge of accurate verification becomes pronounced due to variations in writing styles, distortions, and limited labeled data. In this paper, we propose a novel Siamese-Transformer network tailored for handwritten signature verification using few-shot learning. By synergizing Siamese neural networks and Transformer architectures, our model excels in capturing contextual relationships and discerning genuine from forged signatures. A triplet loss function facilitates discriminative feature learning. Convolution layers extract local features from an image, while the transformer component utilizes these local features to capture global dependencies within signatures. Experimental results on benchmark datasets showcase the model's superior performance in few-shot verification scenarios, marking it as a promising advancement in signature verification and few-shot learning techniques.
ICCIT, 2023
Understanding the behavior and spread of cancer cells within tissues depends heavily on spatial k... more Understanding the behavior and spread of cancer cells within tissues depends heavily on spatial knowledge. Largescale high-resolution cellular data collecting made possible by advanced imaging techniques has created new prospects for detailed examination and characterization of malignant tissues. This study uses segmentation, a technique for separating and identifying individual cells from complicated tissue pictures, to investigate the spatial awareness of cancer cells. Modern deep learning algorithms are used in the suggested segmentation method to automatically identify and separate cancer cells from microscope pictures. The spatial relationships between nearby cells and their surrounding milieu can be evaluated by isolating individual cells and defining their boundaries. This method provides important insights into tumor development patterns, cell interactions, and potential spreading tendencies. Our motive is to investigate the application of spatial analysis to identify spatial clusters, patterns of cellular distribution and problems related tumor.These findings may be important for defining tumor heterogeneity, identifying cancer stem cell niches, and understanding how spatial factors influence therapy response and resistance.
ICCIT, 2023
Bone metastasis is a frequently occurring disease
and can be a consequence of a number of differe... more Bone metastasis is a frequently occurring disease
and can be a consequence of a number of different cancers
such as - prostrate, lung and breast cancers, and predicting
them can be really useful for the diagnosis of patients with such
diseases. Classifying images of bone scan for bone metastasis
prediction requires a huge amount of data to produce a prediction
output which is reliable and accurate, but a single medical
organization usually do not have access to such amounts of data
from other organizations and those organizations are not also
ready to share their patients’ private data as well due to data
security issues. For such scenarios, it is not often possible to
train a model with enough data, thus leading to an inaccurate
prediction model for bone metastasis. This can be devastating
at times due to the occurrence of many false positives or false
negatives, if bone metastasis is wrongly classified. In order to
find a better solution, so that there is less data protection and
privacy issues and therefore more availability of data, we are
proposing to use a Federated Learning (FL) based approach for
bone metastasis prediction using convolutional neural network.
As per our knowledge and background study, we are the first to
use federated learning for bone metastasis prediction on the BS80K dataset. Federated Averaging (FedAvg) strategy was used for
implementing the federated learning methodology where different
client models were built along with a Global Model.
2023 Computer Applications & Technological Solutions (CATS), 2024
Social media plays a vital role in our daily lives. To understand and interpret emotions and opin... more Social media plays a vital role in our daily lives. To understand and interpret emotions and opinions expressed on social media platforms, analyzing sentiment is very important. Our study is based on Twitter sentiment analysis. Our aim is to classify tweets automatically as positive, negative, or neutral based on their content using natural language processing and machine learning algorithms. The dataset we used for our analysis is extracted from the website called mendeley data and also we have added some tweets manually which covers various topics. To remove noise, including URLs, hashtags, punctuations, and user mentions, and to retain essential textual content and emojis, we pre-processed the dataset. Additionally, for our research, we used VADER (Valence Aware Dictionary and sentiment Reasoner) and Transformers-RoBERTa to analyze the sentiment of various tweets. We evaluate the performance of these two models using evaluation metrics such as accuracy, precision, recall and F1-score, and also confusion metrics on the testing set. We also discuss the study's limitations and conclude that machine learning-based sentiment analysis models are a reliable tool for the sentiment analysis of the twitter dataset.
2023 Computer Applications & Technological Solutions (CATS), 2024
A computer programme that imitates and processes human interaction, either through the use of voi... more A computer programme that imitates and processes human interaction, either through the use of voice or text communication, is known as a chatbot. Its purpose is to be of assistance in the process of finding a solution to a problem. The transformation brought on by advances in technology has had an effect on every industry. The chatbot provides assistance with a wide variety of tasks, including Reservations, Customer Service, and a great number of other services. The fast development of technologies relating to artificial intelligence and natural language processing has resulted in an increase in the use of chatbots in a variety of fields, most notably in the field of customer service. Customers could receive advice that is prompt, accurate, and personalised through the use of chatbots, which has the potential to completely transform customer service. Because it can automate customer service and reduce the amount of work that needs to be done by humans, it has gained a lot of popularity in the business world. Which can help businesses improve the experience they provide for their customers. The purpose of this research is to undertake a comparative review of customer service chatbots, with a particular emphasis on their efficiency, usability, and application across a variety of business sectors. The research will uncover best practises, difficulties, and potential for improvement by analysing a variety of chatbot solutions.
2023 Computer Applications & Technological Solutions (CATS), 2024
In the age of information overload, where the world is getting increasingly digital, traditional ... more In the age of information overload, where the world is getting increasingly digital, traditional methods of learning are becoming tedious and extensively outdated. Imagine a system with just a few clicks that can quickly generate perplexing questions and enlightening solutions from a given text. Likewise, this paper represents a groundbreaking system that uses stateof-the-art of natural language processing techniques to analyze subject-specific chapters to create questions and corresponding solutions of varying lengths. The system's versatility as an ideal tool for a wide range of users, including students, researchers, and educators, is a result of its capability to handle a wide range of domains. By offering questions of insight along with appropriate answers, this aforementioned structure demonstrated its extraordinary accuracy and competency in our studies on an array of informational datasets. Altering the learning process and promoting knowledge discovery, this program is flexible in delivering brief or comprehensible solutions to inquiries, having the potential to completely change how individuals interact with written material, whether they are reading for short reference or conducting any depth research. Our suggested framework establishes a dynamic platform for immediate information that enables people to learn substantially more and comprehend any topic in depth. With this leading-edge method, bid goodbye to exhausting manual question generation and get ready to embrace a new era of seamless and fruitful learning with this cutting-edge system.
14th International Conference on Computing Communication and Networking Technologies (ICCCNT), 2024
Because of recent improvements to Bangladesh's roads and highways, Automatic Number Plate Recogni... more Because of recent improvements to Bangladesh's roads and highways, Automatic Number Plate Recognition (ALPR) has become a crucial component. Numerous crimes, including kidnapping, failure to pay tolls, and harassment of women, occur on both public and private transportation. The security forces will be able to locate offenders more quickly with the earlier and more accurate detection of license plates. The authors of this research proposing a deep learning-based fusion model for ALPR that integrates CNN and GRU on the basis of these circumstances. A total of 4753 images from various Bangladeshi roads and highways have been collected for training, validation, and testing purposes. The dataset consists of three classes of data namely Private cars, Public buses, and Trucks where all the images are in RGB format. To get precise and reliable findings, a variety of preprocessing approaches have been applied. After passing the images to the proposed architecture all the necessary parameters have been fine-tuned that causes a lesser amount of trainable parameters and more accuracy. The research demonstrates that the suggested CNN-GRU based fusion architecture, with a 98.97% F1-score, outperforms the leading models. Both static photos and CCTV video material can be used to accomplish ALPR tasks with comparable efficiency. Later, Explainable Artificial Intelligence (XAI) model SHAP has been used in order to interpret the outstanding result with a region of features.
Computing and Informatics , 2024
The heart of any substantial search engine is a crawler. A crawler is a program that collects web... more The heart of any substantial search engine is a crawler. A crawler is a program that collects web pages by following links from one web page to the next. Due to our complete dependence on search engines for finding information and insights into every aspect of human endeavors, from finding cat videos to the deep mysteries of the universe, we tend to overlook the enormous complexities of today's search engines powered by the web crawlers to index and aggregate everything found on the internet. The sheer scale and technological innovation that enabled the vast body of knowledge on the internet to be indexed and easily accessible upon queries is constantly evolving. In this paper, we look at the current state of the massive apparatus of crawling the internet, specifically focusing on deep web crawling, given the explosion of information behind an interface that cannot be extracted from raw text. We also explore distributed search engines and the way forward for finding information in the age of large language models like ChatGPT or Bard. Our primary goal is to explore the junction of large-scale web crawling and search engines in an integrative approach to identify the emerging challenges and scopes in massive data where recent advancements in AI upend traditional means of information retrieval. Finally, we present the design of a new asynchronous crawler that can extract information from any domain into a structured format.
20th International Conference on Informatics in Control, Automation and Robotics, 2023
The study proposes a distributed machine learning-based university recommendation system (URS) in... more The study proposes a distributed machine learning-based university recommendation system (URS) in Bangladesh to help undergraduate students make informed decisions based on user ratings. The system uses advanced distributed machine learning models such as collaborative filtering and popularity-based recommender model which consists of KNNwithmeans model and singular value decomposition (SVD) model to process data and provide accurate recommendations, significantly enhancing the university selection process for students. This study advances educational technology and provides a useful tool for undergraduates in Bangladesh.
2022 IEEE International Conference on Artificial Intelligence in Engineering and Technology (IICAIET), Sep 13, 2022
Springer eBooks, Dec 13, 2022
2022 International Conference on Engineering and Emerging Technologies (ICEET), Oct 27, 2022
Polycystic Ovary Syndrome (PCOS) is an endrocrinological dysfunction prevalent among women of rep... more Polycystic Ovary Syndrome (PCOS) is an endrocrinological dysfunction prevalent among women of reproductive age. PCOS is a combination of syndromes caused by an excess of androgens-a group of sex hormones-in women. Syndromes including acne, alopecia, hirsutism, hyperandrogenaemia, oligoovulation, etc. are caused by PCOS. It is also a major cause of female infertility. An estimated 15% of reproductive-aged women are affected by PCOS globally. The necessity of detecting PCOS early due to the severity of its deleterious effects cannot be overstated. In this paper, we have developed PCONet-a Convolutional Neural Network (CNN)-to detect polycistic ovary from ovarian ultrasound images. We have also fine tuned Incep-tionV3-a pretrained convolutional neural network of 45 layers-by utilizing the transfer learning method to classify polcystic ovarian ultrasound images. We have compared these two models on various quantitative performance evaluation parameters and demonstrated that PCONet is the superior one among these two with an accuracy of 98.12%, whereas the fine tuned InceptionV3 showcased an accuracy of 96.56% on test images.
This paper, deals with systematic study of simple segmentation and classification algorithms for ... more This paper, deals with systematic study of simple segmentation and classification algorithms for kidney tumor using Computed Tomography images. Tumors are of different types having different characteristics and also have different treatment. It becomes very important to detect the tumor and classify it at the early stage so that appropriate treatment can be planned. This CT scans are visually examined by the physician for detection and diagnosis of kidney tumor. However this method lacks accuracy and detection of size of the tumor. So to overcome this, a computer aided segmentation technique has been proposed, which extracts the tumor part from the kidney, further on which feature extraction method is performed for extracting certain features and the type of tumor i.e. malignant or benign is displayed by using simple classifiers .
Uploads
Papers by Md Humaion Kabir Mehedi
and can be a consequence of a number of different cancers
such as - prostrate, lung and breast cancers, and predicting
them can be really useful for the diagnosis of patients with such
diseases. Classifying images of bone scan for bone metastasis
prediction requires a huge amount of data to produce a prediction
output which is reliable and accurate, but a single medical
organization usually do not have access to such amounts of data
from other organizations and those organizations are not also
ready to share their patients’ private data as well due to data
security issues. For such scenarios, it is not often possible to
train a model with enough data, thus leading to an inaccurate
prediction model for bone metastasis. This can be devastating
at times due to the occurrence of many false positives or false
negatives, if bone metastasis is wrongly classified. In order to
find a better solution, so that there is less data protection and
privacy issues and therefore more availability of data, we are
proposing to use a Federated Learning (FL) based approach for
bone metastasis prediction using convolutional neural network.
As per our knowledge and background study, we are the first to
use federated learning for bone metastasis prediction on the BS80K dataset. Federated Averaging (FedAvg) strategy was used for
implementing the federated learning methodology where different
client models were built along with a Global Model.
and can be a consequence of a number of different cancers
such as - prostrate, lung and breast cancers, and predicting
them can be really useful for the diagnosis of patients with such
diseases. Classifying images of bone scan for bone metastasis
prediction requires a huge amount of data to produce a prediction
output which is reliable and accurate, but a single medical
organization usually do not have access to such amounts of data
from other organizations and those organizations are not also
ready to share their patients’ private data as well due to data
security issues. For such scenarios, it is not often possible to
train a model with enough data, thus leading to an inaccurate
prediction model for bone metastasis. This can be devastating
at times due to the occurrence of many false positives or false
negatives, if bone metastasis is wrongly classified. In order to
find a better solution, so that there is less data protection and
privacy issues and therefore more availability of data, we are
proposing to use a Federated Learning (FL) based approach for
bone metastasis prediction using convolutional neural network.
As per our knowledge and background study, we are the first to
use federated learning for bone metastasis prediction on the BS80K dataset. Federated Averaging (FedAvg) strategy was used for
implementing the federated learning methodology where different
client models were built along with a Global Model.