Browsing by Author "Mamun, Shamim Al"

Now showing 1 - 8 of 8

A deep convolutional neural networks for brain tumor classification based on MRI images
(Daffodil International University, 2024-01-01) Mamun, Shamim Al
The tenth most common cause of death for both men and women is cancer of thebrain and associated nervous system tumors. According to estimates, primarymalignant brain and central nervous system tumors will be the cause of 18,980fatalities in the United States in 2023 (11,020 males and 7,970 women). Primarymalignant brain and central nervous system tumors are expected to have killed251,329 persons globally in 2020. So the classification of brain tumor is reallyimportant because identification of disease is the first step to prevent that. For that I are used to classify brain tumor using deep learning.For classification of brain tumor I applied three deep learning model one ml framework and two ml domain. Like Vgg16, ResNet50, Xception ,PyTorch , transfer learning and fine-tuning. After applyingon4278 MRI images I got loIst accuracy is 83.85% and highest accuracy is 98.01%inPyTorch fine-tuning.
ADVANCEMENT OF BASE CASE TO THE RECURSION: A WAY TO IMPROVE THE RUNNING TIME OF THE TRADITIONAL MERGE SORT ALGORITHM
(Daffodil International University, 2007-07-01) Talukder, Ashis; Mamun, Shamim Al; Alam, Abu Md. Zafor
Merge sort is one of the best examples of divide and conquer technique. Divide and conquer is inherently recursive by nature. So if the number of recursive calls can be decreased, the complexity of any recursive algorithm can be improved. This naturally leads to the advancement of base case of recursion. In this research paper, we propose such a base case-advancement technique. We will apply it upon the traditional merge sort and show that it improves the number of recursive calls, the running time as well as the memory requirement.
Cascade Classification of Face Liveliness Detection Using Heart Beat Measurement
(Scopus, 2021) Rahman, Md. Mahfujur; Mamun, Shamim Al; Kaiser, M. Shamim; Islam, Md. Shahidul; Rahman, Md. Arifur
Face detection and recognition is a prevalent concept in security and access control area which is commonly used in surveillance cameras at public places, attendance etc. But often this type of system can be circumvented by holding a photo or running a video of authorized person to the camera. Therefore, liveliness concept comes up with a solution to detect the person is real or spoofed. In this paper, we proposed a cascade classifier based model for detecting liveliness using deep-learning and Heart-beat measurement. Moreover, we have evaluated our model accuracy with our own dataset of real and fake videos and photos. By using our proposed model of face liveliness detection model, FPR and FNR have declined 16% and 5.22% respectively. In addition, we have also compared proposed system with other state-of-art methods. And here proposed study has achieved an accuracy of 99.46%.
Explainable Mortality Prediction Model for Congestive Heart Failure with Nature-Based Feature Selection Method
(MDPI Publications, 2023-05-17) Tasnim, Nusrat; Mamun, Shamim Al; Islam, Mohammad Shahidul; Kaiser, M. Shamim; Mahmud, Mufti
A mortality prediction model can be a great tool to assist physicians in decision making in the intensive care unit (ICU) in order to ensure optimal allocation of ICU resources according to the patient’s health conditions. The entire world witnessed a severe ICU patient capacity crisis a few years ago during the COVID-19 pandemic. Various widely utilized machine learning (ML) models in this research field can provide poor performance due to a lack of proper feature selection. Despite the fact that nature-based algorithms in other sectors perform well for feature selection, no comparative study on the performance of nature-based algorithms in feature selection has been conducted in the ICU mortality prediction field. Therefore, in this research, a comparison of the performance of ML models with and without feature selection was performed. In addition, explainable artificial intelligence (AI) was used to examine the contribution of features to the decision-making process. Explainable AI focuses on establishing transparency and traceability for statistical black-box machine learning techniques. Explainable AI is essential in the medical industry to foster public confidence and trust in machine learning model predictions. Three nature-based algorithms, namely the flower pollination algorithm (FPA), particle swarm algorithm (PSO), and genetic algorithm (GA), were used in this study. For the classification job, the most widely used and diversified classifiers from the literature were used, including logistic regression (LR), decision tree (DT) classifier, the gradient boosting (GB) algorithm, and the random forest (RF) algorithm. The Medical Information Mart for Intensive Care III (MIMIC-III) dataset was used to collect data on heart failure patients. On the MIMIC-III dataset, it was discovered that feature selection significantly improved the performance of the described ML models. Without applying any feature selection process on the MIMIC-III heart failure patient dataset, the accuracy of the four mentioned ML models, namely LR, DT, RF, and GB was 69.9%, 82.5%, 90.6%, and 91.0%, respectively, whereas with feature selection in combination with the FPA, the accuracy increased to 71.6%, 84.8%, 92.8%, and 91.1%, respectively, for the same dataset. Again, the FPA showed the highest area under the receiver operating characteristic (AUROC) value of 83.0% with the RF algorithm among all other algorithms utilized in this study. Thus, it can be concluded that the use of feature selection with FPA has a profound impact on the outcome of ML models. Shapley additive explanation (SHAP) was used in this study to interpret the ML models. SHAP was used in this study because it offers mathematical assurances for the precision and consistency of explanations. It is trustworthy and suitable for both local and global explanations. It was found that the features that were selected by SHAP as most important were also most common with the features selected by the FPA. Therefore, we hope that this study will help physicians to predict ICU mortality for heart failure patients with a limited number of features and with high accuracy.
HealthBlock: A Secured Healthcare System Using Blockchain
(Springer Nature, 2023-06-03) Rahman, Md. Mahfujur; Sifat, Md. Nur Amin; Rahman, Mostafizur; Rahman, Mushfiqur; Mamun, Shamim Al; Kaise, M. Shamim
Blockchain technology enables a distributed and decentralized environment with no more central authority. To increase the accuracy of electronic healthcare records (EHRs) and establish a secured patient-centric approach, Blockchain can be a smart solution in this case. Blockchain is a distributed ledger technology that allows for the secured transfer of medical records while a transaction is created in the network. As Blockchain is decentralized and transparent if one user tampers with medical records of transactions, all other nodes/participants would cross-reference each other and easily pinpoint the node/participants with the wrong information. In this paper, we proposed a smart contact-based Blockchain technology and implemented it in the Hyperledger framework to make the healthcare system more confidential and secure. Moreover, Smart contracts give us more security while transacting data and reduce the entire transaction cost, and also make faster transaction speed.
Implementation of Real-Time Automated Attendance System Using Deep Learning
(Daffodil International University, 2022-02-05) Hasan, Hafiz Mahdi; Rahman, Md. Mahfujur; Khan, Md. Al-Amin; Meghla, Tamara Islam; Mamun, Shamim Al; Kaiser, Shamim
In comparison to general manual operations, contemporary technology always saves time and is often more hassle-free when it comes to verifying human authenticity using their biometrical components. However, despite the fact that face recognition technology has been used in a variety of sectors such as human identification systems, this work is the first to describe how the Face Recognition Technique can be integrated with a deep learning approach. Advanced deep learning techniques can make the attendance system completely automated, highly secure, easier to use, and faster to implement than older systems. Nowadays, the Attendance System is becoming increasingly automated, resulting in time-saving, effective, and beneficial solutions that reduce the burden on administration and organizations. In this paper, we suggest an automatic attendance mechanism that is based on Deep Convolutional Neural Networks (DCNN). SeetaFace, a deep convolutional neural network-based face detection system, is employed in this research effort to detect faces in real-time video capture. This implementation is a VIPLFaceNet implementation, to be more specific. AlexNet, which is also a DCNN, is used for image categorization. The experimental results bring four short similarity situations of the classroom such as absence, delayed appearances, early leave, and unauthorized entry during class or session along with the name, student id, and section and passes this information to the attendance sheet which will evaluate the students/persons in the classroom. This methodology saves time when compared to the traditional method of attendance marking, as well as allows organizations to conduct stress-free observations of students and staff.
Performance Analysis of Machine Learning Approaches In Software Complexity Prediction
(Scopus, 2021) Reza, Sayed Moshin; Rahman, Md. Mahfujur; Parvez, Hasnat; Badreddin, Omar; Mamun, Shamim Al
Software design is one of the core concepts in software engineering. This covers insights and intuitions of software evolution, reliability, and maintainability. Effective software design facilitates software reliability and better quality management during development which reduces software development cost. Therefore, it is required to detect and maintain these issues earlier. Class complexity is one of the ways of detecting software quality. The objective of this paper is to predict class complexity from source code metrics using machine learning (ML) approaches and compare the performance of the approaches. In order to do that, we collect ten popular and quality maintained open source repositories and extract 18 source code metrics that relate to complexity for class-level analysis. First, we apply statistical correlation to find out the source code metrics that impact most on class complexity. Second, we apply five alternative ML techniques to build complexity predictors and compare the performances. The results report that the following source code metrics: Depth inheritance tree (DIT), response for class (RFC), weighted method count (WMC), lines of code (LOC), and coupling between objects (CBO) have the most impact on class complexity. Also, we evaluate the performance of the techniques, and results show that random forest (RF) significantly improves accuracy without providing additional false negative or false positive that work as false alarms in complexity prediction.
Simulation and comparative study of different feedback-based matchline sensing circuits for Ternary Content Addressable Memory(TCAM)
(Department of Electrical and Electronic Engineering, Islamic University of Technology (IUT) The Organization of Islamic Cooperation (OIC) Board Bazar, Gazipur-1704, Bangladesh, 2021-03-30) Hasan, Abir; Mamun, Shamim Al; Soumik, Sadman Sakib
Content addressable memory (CAM) can perform high-speed table look-up with bit level masking capability. The search and detection of data stored in CAM cells are done using different sensing circuitry. Current Race schemes serves this purpose with great efficiency. Adding feedback system in the circuitry increases the performance and energy efficiency. In this thesis, a comparative study has been conducted among different CR schemes and the simulation profiles has been observed based on a 16 by 32 Ternary CAM (TCAM) array