2022

Browse

Now showing 1 - 20 of 52

3D Object Detection with Stereo Vision and Transformer
(Department of Computer Science and Engineering(CSE), Islamic University of Technology(IUT), Board Bazar, Gazipur, Bangladesh, 2022-05-30) Samin, Abid Ahsan; Hassan, Abdullah; Khan, Md. Rakib Hossain
Identifying 3D objects with computer vision in a precise manner has been a challenging task in the field of autonomous driving. Partly because it requires proper depth estimation. Until now, Li-DAR technology has been used to achieve this task which is precise but also expensive. The introduction of pseudo Li-DAR promises an alternative approach which is cheaper with fairly good precision. However, pseudo Li-DAR can be replaced with 2D image representation with similar precision. Transformer is another technology which is widely used to process sequential data. Recent studies show that transformer can also be used for object detection purposes. In this literature, we look into the concept of pseudo Li-DAR, image representation of depth and detection transformer(DETR). Later, we introduce a new approach of using image based depth output with DETR to achieve accurate object detection. Finally, we compare our results with other available methods used for object detection in order to establish a benchmark.
A Benchmark for Detection and Recognition of Bangladeshi Traffic Signs in Real-world Images
(Department of Computer Science and Engineering(CSE), Islamic University of Technology(IUT), Board Bazar, Gazipur, Bangladesh, 2022-05-30) Khan, Rizwanul Haque; Shanto, Md Saimul Haque; Ashik, Ahmed Nusayer
Traffic sign detection is an indispensable part of autonomous driving and transportation safety systems. However, the accurate detection and recognition of traffic signs remain challenging, especially under extreme conditions, such as various weather and geo-social features. Though a lot of work has been done in the domain of Traffic Sign Detection and Recognition (TSDR) systems, only a few of them focus on a dataset that comprises the real-world challenges. Moreover, in the context of Bangladesh, there is no well-defined public dataset, let alone one that focuses on real-life challenges. The geo-social features of Bangladesh add some unique challenges that are not seen in most parts of the world. This proposal aims to address the lack of quality Bangla traffic sign detection dataset. To accomplish this task, traffic sign images will be extracted from videos collected from Bangladeshi roads. A performance benchmark will be presented by applying state-of-the-art methods to the said dataset. Using the best-performing method, an autonomous driver notification system will be developed to alert the drivers on the go.
A Comparative Study of Static Code Metrics and Behavioural Metrics for Predicting Risk Scores in Android Apps
(Department of Computer Science and Engineering(CSE), Islamic University of Technology(IUT), Board Bazar, Gazipur, Bangladesh, 2022-05-30) Nafis, Fahim Arsad; Munia, Maysha Afrin; Saiara, Syeda Mishra
Identifying security flaws and distinguishing non-susceptible code from vulnerable code is a difficult undertaking. Security flaws are usually inert until they are exploited. Software metrics have been widely utilized to forecast and signal a variety of software quality features. We investigate static code metrics and behavioral code metrics, their correlation, and their association with security vulnerabilities in Android applications. The aim of the study is to understand: (i) the comparison between static software metrics and behavioral code metrics; (ii) the ability of these metrics to predict security vulnerabilities, and (iii) which are the strongly correlated static code metrics and behavioral code metrics. From our study, we have found that even though static code metrics require higher computational power, it provides better results to predict the risky behavior of android applications and Random Forest Regression provides more stable results with a better R2 score for this specified dataset which we create for our thesis.
A Hybrid Blockchain Architecture for Cross-Platform Transaction
(Department of Computer Science and Engineering(CSE), Islamic University of Technology(IUT), Board Bazar, Gazipur, Bangladesh, 2022-05-31) Meem, Jarin Tanzim; Yasmin, Khan Silvi; Ruhee, Rokeya Samantha
Blockchain has emerged as one of the most significant technological breakthroughs in software design and technology over the past ten years. Existing blockchain architectures have two major limitations: first, the blockchain itself is vulnerable to attacks such as double spending, selfish mining and IoT smart devices, and distributed denial of service attacks, are also vulnerable once hackers successfully infiltrate blockchain systems; second, because IoT devices are heterogeneous and have resource limitations, implementation of current blockchain systems in the IoT scenario cannot reflect strong adaptability and meet IoT s requirements; and third, because IoT devices are heterogeneous and have resource constraints Our research focuses on the development of a hybrid architecture that facilitates the exchange of assets across multiple blockchains while also improving the privacy of the current blockchain by ensuring data integrity and data integrity fidelity.
A model agnostic explainable approach for detecting Cyber bullying in Bangla language using transformer based models
(Department of Computer Science and Engineering(CSE), Islamic University of Technology(IUT), Board Bazar, Gazipur-1704, Bangladesh, 2022-05-30) Nobo, Takia Mosharref; Galib, Mostafa; Rabib, Hasnain Karim
Almost every facet of social communication has changed as a result of the exponential growth of social media platforms usage. Meanwhile, evidence is accumulating that the rising usage of social networks in the digital realm has given rise to an unsettling problem that has resurrected in new contexts: cyberbullying. The majority of current cyberbullying detection research focuses on English texts. On the other hand, while being spoken by 230 million people globally and being rich in diversity, the Bengali language is under-resourced for natural language processing (NLP). Recently, there has been an alarming surge in the number of incidences of gender-based discrimination or sexual harassment expressed on social media sites. In this study, we presented the cyberbullying detection under different categories in low-resourced Bangla language using transformer based models. We created our own dataset on gender discrimination and appended it to another open-source Bangla dataset with 4 classes. In our proposed approach, we used five different models to train our augmented dataset, followed by an ensembling technique on those five models.Then we make the models explainable using model agnostic approaches. Finally, we compared the individual prediction accuracies with the ensembled prediction accuracies. While training the dataset, we followed the stratified k-fold cross validation technique. Our evaluations yield up to an Accuracy of 75% in cyberbullying detection on emsembling.
A Semi-Automated Approach to Generate Bangla Dataset for Question-Answering and Query-Based Text Summarization
(Department of Computer Science and Engineering(CSE), Islamic University of Technology(IUT), Board Bazar, Gazipur, Bangladesh, 2022-05-30) Mushabbir, Mueeze Al; Alamgir, Refaat Mohammad; Humdoon, Ahmed Azaz
With the vast amount of information available on the Internet, finding answers to questions is as important as ever in today’s day and age. In Natural Language Processing Research, Question Answering (QA) and Query-based Text Summarization (QBSUM) are there to tackle this challenge. However, most of the work being done neglects low resource languages such as Bangla, resulting in the small number of quality datasets available in the literature. Therefore to address this research gap, in this work, we propose a semi-automated methodology for generating a Bangla dataset with Natural Questions for three tasks - Question Answering (QA), Query-based Single Document Text Summarization (SD-QBSUM) and Query-based Multi-Document Text Summarization (MD-QBSUM). We then provide baselines for this dataset on those tasks and also compare our dataset with existing ones on various metrics.
A Study of Permission-based Malware Detection Using Machine Learning
(Department of Computer Science and Engineering(CSE), Islamic University of Technology(IUT), Board Bazar, Gazipur, Bangladesh, 2022-05-31) Islam, Md. Rafid; Rahman, Ratun; Ahmed, Akib
Malware is becoming more prevalent, and several threat categories have risen dramatically in recent years. This paper provides a bird's-eye view of the world of malware analysis. It also presents a brief review of malware analysis approaches, common detection types, and some basic preventive strategies from various angles. An experiment has been done to show the influence of human factors on people. This study shows that most people are more likely to fall victim to a malware attack if that seems to come from a reliable source or person. The efficiency of five different machine learning methods (Naive Bayes, K-Nearest Neighbor, Decision Tree, Random Forest, Decision Forest) combined with features picked from the retrieval of Android permissions to categorize applications as harmful or benign is investigated in this study. On a test set consisting of 1,168 samples (each consisting of 948 features), produce accuracy rates above 80% (Except Naive Bayes Algorithm with 65% accuracy). Of the considered algorithms TensorFlow Decision Forest performed the best with an accuracy of 90%.
An Empirical Study on Neophytes of Stack Overflow: How Welcoming the Community is Towards Them
(Department of Computer Science and Engineering(CSE), Islamic University of Technology(IUT), Board Bazar, Gazipur, Bangladesh, 2022-05-30) Mohammad, Suzad; Jobair, Abdullah Al; Maisha, Zahin Raidah
Stack Overflow (SO) is the most popular question and answers (Q&A) platform for programmers with a rapidly expanding community of new users. However, the unwelcoming environment towards new users has been under discussion for several years which is a major concern and hindrance towards the enhancement of a skillful community. In this work, we study a specific group of users who are either registered in the last 45 days or have a reputation less than or equal to 50 and term them as "neophytes". Upon establishing significance of the definition of neophytes, we perform manual analysis of neophytes' posts. We organize our research work into two research questions where we investigate whether neophytes actually face hurdles while collaborating in Stack Overflow and, if so, identify the potential reasons behind this phenomenon by qualitative and quantitative analysis. Our study finds that neophytes are indeed facing hurdles while collaborating in the platform. The reasons behind the hurdles include harsh moderation of posts, negligence of the posts, deleting or closing of posts, downvoting without providing any proper reasoning, etc. Our findings can provide guidelines to create a more user-friendly SO community. Furthermore, this study can guide researchers to observe the reactions of neophytes in adverse situations and recommend some steps for the community to make positive changes to the Stack Overflow environment.
An End to End System for Online Handwritten Bangla Character Recognition
(Department of Computer Science and Engineering(CSE), Islamic University of Technology(IUT), Board Bazar, Gazipur, Bangladesh, 2022-05-30) Nahin, Shahriar Nur; Imam, Kazi Fahim; Rahman, Nabil; Tasnim, Anika
This report summarizes the attempt to find the way towards building an Optical Character Recognition System for handwritten Bangla characters. The complex and unique structure of scripts like Bangla and ever challenging nature of hand- written texts combined makes it really difficult to complete a perfect system to approach to convert the scanned handwritten Bangla scripts to machine editable digital counterpart format of it- as segmentation of the whole image into char- acters and then classification of the segmented characters is difficult enough to make the task challenging. In our work, we propose to approach the segmentation process (directly segment to words) with Distance Transform and morphological operations for error correction later. Then two zone approach (either side of matra- upper and lower zone) and apply connected component analysis on both zones. We handled or adjusted the failed and not directly successful cases by experimenting with the characteristics of handwritten characters. Then for clas- sification process, we proposed to classify the segmented characters using neural networks trained on the relatively newly available datasets. Multiple column, Mixed characters (Bangla- other languages) and Scene Text Recognition is out of the scope of our study so far. And we could not include the post-processing part for our work for lack of work or mention in existing literature, which might be a great addition in the way of building a complete OCR system.
An Exploratory Study of DevOps Approach and Engagement : From the Perspective of Bangladesh IT Industries
(Department of Computer Science and Engineering(CSE), Islamic University of Technology(IUT), Board Bazar, Gazipur, Bangladesh, 2022-05-30) Emad, Mahir Ashraf; Evan, H.M.Raihan; Azad, Anirban Tasfin
On-Time Delivery (OTD) with high quality is the key success factor for Software development organizations. To increase the ability to meet the OTD with high quality, software organizations start practicing DevOps. And the rate of adopting DevOps is highly observed in the regions where the IT industry is already developed. Still, there is a clear demand of analyzing the DevOps culture to understand the formulation and vision of continuous quality product delivery of Software organizations in a region like Bangladesh where the IT sector is rapidly expanding. In order to explore this, we conduct our current study using both empirical study and statistical analysis based on semi-structured interviews and surveys. We conduct our survey on 9 different Bangladeshi software organizations ranging in size from small to large. It is discovered that most of the organizations follow a traditional approach, rather than having an established DevOps team. A large number of developers are accountable for the DevOps pipeline and the surveyed organizations in this region had a positive mentality about adopting DevOps.
An Indoor Object Dataset for Mobile-based Detection and Recognition Systems for the Visually Impaired
(Department of Computer Science and Engineering(CSE), Islamic University of Technology(IUT), Board Bazar, Gazipur, Bangladesh, 2022-05-30) Azad, Shehreen; Sayed, Abdullah Abu; Faiyrooz, Noshin
Indoor object detection is a challenging area of computer vision where comparatively lesser work has been done compared to its outdoor counterpart. Surely, such a task requires huge amount of training data to make any classifier detect indoor objects with high precision. This indoor object detection can become way more challenging when it has to be specifically tailored for visually impaired people’s mobility and interaction with everyday use objects. This report presents a novel indoor object dataset containing 5196 unique images of 8 everyday use indoor object category relevant to daily interaction of visually impaired people. The uniqueness of this dataset compared to existing indoor objects dataset is this dataset deals with everyday use objects and presents them with more contextual information than that is available in existing literature. Moreover, the varying lighting condition, non-canonical viewpoints, occlusion and complex background makes the dataset more robust while being trained on any object detection algorithm. Instead of going for higher accuracy we aim to find a trade-off between accuracy and speed as if this dataset is to be used to build a system for visually impaired people’s navigation needs, that system has to be deployed on mobile or sensor-based hand-held device which requires lightweight models. Hence our proposed dataset is tested on two light-weight model, namely, SSD MobileNet V2 FPNLite and EfficientDet D0. It has achieved a mean average precision (mAP) of 29.5 and 39.4 respectively on both the models which is better than the original mAP values achieved by these models. Our proposed dataset can be extended with other indoor object detection dataset, as well as it can be used to build a system for visually impaired people’s navigation.
Analyzing Web Application Vulnerabilities of Educational Institutions in Bangladesh
(Department of Computer Science and Engineering(CSE), Islamic University of Technology(IUT), Board Bazar, Gazipur, Bangladesh, 2022-05-31) Karim, Mahbul; Muntakim, Afia; Zaman, Hridita Nur
The biggest challenge we face today is web security. It is the fundamental framework for the global data society. People’s daily activities mostly depend on internet-based applications. No web application is free from threats and security issues. Clients and users make mistakes when interacting with web applications, which can lead to security issues. Besides, there are coding flaws and server misconfiguration issues which gradually lead to service failure or attacks on vulnerable information. Strong security in the web application is a vital need for online presence nowadays. Dealing with web security issues requires deep insight as there are a lot of tools available to detect vulnerabilities. Proper understanding and deep analysis are required to find the proper tool for this application. This study aims to detect vulnerabilities of the educational websites in Bangladesh and analyze which scanning tool provides more accurate results. For our analysis, we have used the two most prominent web application security scanners, Acunetix and Nikto. After scanning, many security issues and vulnerabilities were found. However, the most common vulnerability issue among all the websites was SQL injection, XSS, and Clickjacking
Automated Sleep Scoring Using Multichannel EEG Biomarkers
(Department of Computer Science and Engineering(CSE), Islamic University of Technology(IUT), Board Bazar, Gazipur-1704, Bangladesh, 2022-05-30) Bari, Md. Abdul; Jany, Rafsan; Uddin, Musfik
Sleep staging is one of the most essential approaches for diagnosing many sorts of sleep- related illnesses. Electroencephalography (EEG) is considered a computing tool for eval- uating the relationship between neurological effects and sleep stages because it detects sleep-related neurological changes quickly and accurately. So In comparison to the tradi- tional polysomnographic signal based approach, EEG is considered to be a more efficient tool to predict sleep stages outside of a fully equipped medical environment. The goal of this study is to use sleep EEG data to identify effective neurological EEG biomark- ers and predict five stages of sleep. We analyzed three EEG channels (F4, C4 and O2) from the dataset collected by Haaglanden Medisch Centrum (HMC, The Netherlands) and published by PhysioNet that contains 154 sleep recordings. In this study we have applied different classification models that are Decision Forest, Support Vector Machine, K-Nearest Neighbors, Extreme Gradient Boost and Neural Network to classify 5-class sleep stages. Among those we found that the Neural Network outperformed other mod- els. We have also identified delta wave power ratios (DAR, DTR, and DTABR) as EEG biomarkers that improved the overall accuracy from 84% to 92% using the Neural Net- work model.
Bioinformatics Analysis of Differentially Expressed Gene's in Breast Cancer Using DESeq2
(Department of Computer Science and Engineering(CSE), Islamic University of Technology(IUT), Board Bazar, Gazipur, Bangladesh, 2022-05-30) Malick, Sow Bocar Amadou; Conteh, Fatoumatta; Sawo, Muhammed
Differential Gene Expression Analysis is a strong tool for determining if genes in two or more sample groups are expressed at significantly different levels. To estimate gene counts and identify deferentially expressed genes, we’ll utilize the DESeq2 software. Also, while determining whether genes are deferentially expressed, we must account for variation in the data. The purpose is to see if differences between groups are substantial for each gene, given the biological differences between biological replicates. Using Normalized to Read Count Data (NRCD) and statistical analysis, DEG analysis was used to find quantitative differences in expression levels between experimental groups. For example; statistical testing is used to decide whether for a given gene and observed difference in read counts is significant. I.e., whether it is greater than what would be expected just due to natural random variation. The analysis requires gene expression values to be compared between sample group types. The goal is to determine which genes are expressed at different levels between conditions. It has become a widely used technology that allows for effective genome-wide relative gene expression quantification, and it is the method of choice for identifying deferentially expressed genes between two or more biological situations of interest. The primary challenges surrounding such DE analysis have been highlighted from the start, and several methodologies and tools have been offered in the relevant literature. One of the most difficult aspects of this study, as with any other statistical research, has been determining the probabilistic model that best fits the data, as well as the model’s optimal parameter estimates. Another significant challenge was the requirement for data normalization in order to appropriately compare two biological situations by analyzing and removing any potential technological and/or biological biases. Last but not least, several research have emphasized the practical requirement to determine the ideal number of biological replicates per condition and the optimal library size. We’ll go over the use of DeSeq2 method as a utilized methodology and tools for DE analysis in this article. The gene outcomes can offer biological insights into processes affected by the conditions. greater than what would be expected just due to natural random variation.
Blockchain for Electronic Health Record
(Department of Computer Science and Engineering(CSE), Islamic University of Technology(IUT), Board Bazar, Gazipur, Bangladesh, 2022-05-30) Farah, Radwan Mohamed; Atiku, Ali Umar
Blockchain may additionally reinvent the technique victim’s electronic fitness statistics are distributed and kept via imparting more secure mechanisms for reallocate peer-to-peer (P2P). That means to help and simplicity the recognize of this give out ledger era, a strong systematic literature overview changed into manage, attending to discover the latest literature on blockchain and care realm and establish current question and open queries, coached way of the enhance of take a look at queries concerning EHR for the duration of a blockchain. Pretty 300 scientific research revealed within the last 10 years have been researched, culminating in the construction of an up-to-date classification, question and unlocked queries recognized, and additionally the foremost important tactics, statistics types, requirements and architectures concerning using blockchain for EHR had been evaluated and referred to.
Blockchain-based Decentralized Source Code Repository Hosting Service with Middleware Approach
(Department of Computer Science and Engineering(CSE), Islamic University of Technology(IUT), Board Bazar, Gazipur, Bangladesh, 2022-05-30) Islam, Md. Tahmid; Munna, Sakibul Islam; Haque, MD. Rafid
Software developers must work together in order to provide a better product. As a result, many software developers use version control systems throughout project development since it helps them manage source code and enables them to keep track of the many versions they’ve worked on. Items are organized, regulated, and directed by this strategy. In spite of the fact that the version control system is generally decentralized, there is no properly defined practical method for remotely maintaining the code that is not centralized. For a distributed network of computers, we describe and explain our solution, which leverages a blockchain and smart contracts to authorize, monitor, and perform version control actions on a code repository. Using our strategy, there is no need to have a centralized authority that is trustworthy. The immutability of the code and the ownership information of the code writer are protected via the blockchain. A network of servers (IPFS) also maintains the security of the code repository and its content. In this system, the code is stored on a distributed network of servers, ensuring its availability and integrity, while a decentralized blockchain ensures ownership information and the immutability of the repository by encrypting information from the codebase’s remote location with a hash that combines the owner’s key and the entire code repository.
ChartSumm: A large scale benchmark for Chart to Text Summarization
(Department of Computer Science and Engineering(CSE), Islamic University of Technology(IUT), Board Bazar, Gazipur, Bangladesh, 2022-05-30) Rahman, Raian; Hasan, Rizvi; Farhad, Abdullah Al
Information visualization such as bar- and line-charts are quite popular for understanding large tabular data. But, interpreting information solely with different visualization techniques can also be difficult due to different reasons like visual impairment or the requirement of prior domain knowledge to understand the chart. Automatic "chart to text summarization" can be promising and effective tool for providing accessibility as well as precised insights of chart data in natural language. In spite of having a good potential, there have not been a lot of works on chart to text summarization making it a low resource task. Scarcity of large scale datasets for chart to text summarization is one of the reason behind this. The human written descriptions in the available dataset also contains information beyond the knowledge of the chart making it difficult for us to have an unbiased evaluation. In our thesis, we propose ChartSumm a large scale dataset for chart to text summarization consisting of 84,363 charts along with their metadata and descriptions. We also propose two test sets: test-e and test-h for evaluating the performance of the trained models available in this domain. Our experiment shows that a T5 model trained on our dataset has achieved BLEU score of 75.72 in test-e set and 64.78 in test-h set. From our analysis we can conclude that large language models like T5 and BART can generate short precised deception from given chart metadata.
Classification of Stack Overflow Questions Based on Difficulty
(Department of Computer Science and Engineering(CSE), Islamic University of Technology(IUT), Board Bazar, Gazipur, Bangladesh, 2022-05-30) Raida, Maliha Noushin; Sristy, Zannatun Naim; Monisha, Sheikh Moonwara Anjum; Ulfat, Nawshin
Technical question answering sites, like Stack Overflow, are gaining enormous attention from the learners and practitioners of specialized fields to exchange their programming knowledge. Question answering on different topics has engaged all levels of programmers. All the developers don’t have the same level of expertise, and the question differs among them in terms of complexity and context. However, the existing approach of Stack Overflow models primarily filters out the questions based on tags, which is inefficient for predicting the difficulty level. Due to the limitation of the process, a large part of these posts fails to attract the attention of appropriate users, resulting in valid questions having no answer or significant delay in response time. Therefore, to address these limitations, we proposed three different supervised models using TF-IDF, Topic Modeling(LDA), and Doc2Vec that build more complicated relationships by extracting context-dependent features between the user and the question. Each of the models builds an informative relationship that helps classify the difficulty of a question. Extensive experiments on different variations of the datasets demonstrate the improved efficacy of our proposed models over contemporary models. The experiments find out that even with limited information, the models performance scores are satisfactory and the Doc2Vec model outperforms the other models under consideration.
Classifying Stack Overflow Questions Quality using SVM
(Department of Computer Science and Engineering(CSE), Islamic University of Technology(IUT), Board Bazar, Gazipur-1704, Bangladesh, 2022-05-30) Atiku, Muhammad Bello; Ismail, Naayif; Alhadji, Mohamadou
This thesis aims to discover indicators of quality, and to use this knowledge to correctly classify the quality of questions and answers from Stack Overflow. The proliferation of technical questions and answers on Q&A websites such as Stack Overflow means there is more information available than ever. However, the ease of publishing such information also tends to mean the quality varies significantly. The job of moderating Stack Overflow is left to the community. Stack Overflow performs some basic quality analysis, but this is an area where improvement would have many benefits to not only Stack Overflow, but many other domains where the quality of text is important.
Consistency of Comments to Source Code: An Empirical Investigation and A Dataset
(Department of Computer Science and Engineering(CSE), Islamic University of Technology(IUT), Board Bazar, Gazipur, Bangladesh, 2022-05-30) Islam, Maksuda; Haque, Ahsanul; Hossen, Md Safayat
Comments in the code are a primary source for system documentation. These are indispensable for the work of software maintainers as a foundation for code traceability, maintenance activities, and the use of the code itself as a library or framework in other projects. However, the quality of the comment has been overlooked for various reasons. Also, comments are doubtful to change with the evolution of the source code. The source code gets updated whenever the changes occur, but the comments are ignored. It leaves a new developer even more confused. So the coherence between the comments and the source code must be ensured and maintained. This paper aims to provide a dataset consisting of code-comment pairs through our research work. We have annotated 9,311 classes and methods of different C\# projects. 4,953 code comment pairs were taken after removing NULL, constructor, and variable. We employed a metric called Bilingual Evaluation Understudy (BLEU) to validate our human-curated dataset. This paper also includes a comparative analysis and discussion between the human-curated annotation and annotation provided by the BLEU score. A modified model from a previous study is also proposed, which obtained an accuracy of 96.56\% using the performance metric AUC-ROC after fitting the model to our annotated 4,953 code-comment pairs. In contrast, the previous model gave 93\% accuracy using a similar performance metric on this same dataset.