Browsing by Author "Mahmud, Asif"

Now showing 1 - 7 of 7

A Descriptive Study on the Present Prescribing Pattern for Children under 5 Years of Age for Different Diseases in Bangladesh at Dhaka Shishu Hospital
(2020) Raka, Sabreena Chowdhury; Chowdhury, Arifa; Mahmud, Asif; Rahman, Arifur
In the present time, the percentage of children fall in sick is relatively higher. For this reason, they get admission into the hospital or has to visit the hospital and get prescribed different medicines from the doctors. For different reasons sometimes those medicines may be irrational. This survey was carried out from Dhaka Shishu Hospital, to evaluate the children prescription pattern in a hospital of Bangladesh to know the actual scenario. About 1100 prescription were collected from different department of this hospital. Out of 1100 prescriptions, around 67% of medicines were antibiotics, 34% antihistamine, 33% NSAID, 7% ionic supplement, anti-ulcerant and antiemetic. The percentage of the practice of polypharmacy was enormous.
A Proposed novel approach to face recognition using CNN
(BRAC University, 2023-09) Mahmud, Asif; Alam, Md.Golam Rabiul
Facial recognition has emerged as a crucial technology with applications spanning from security and surveillance to user authentication and human-computer interaction. This work is a comprehensive study of facial recognition techniques leveraging Convolutional Neural Networks (CNN), which is a class of deep learning models known for their exceptional performance in image processing tasks. The primary objective of my research is to develop a strong facial recognition system capable of accurately identifying individuals across various real-world scenarios and challenges. The thesis begins by providing an overview of the fundamentals of CNN, their architecture, and their relevance in image-based pattern recognition. Then it delves into the pre-processing steps involved in preparing facial images for CNN-based recognition, including data collection, data augmentation, and face detection. Special attention is given to handling occlusion, illumination variations, and pose changes often encountered in real-world environments. The core of my work focuses on the design and implementation of CNN-based facial recognition models. Different CNN architectures are explored, and their performance is evaluated using benchmark datasets. In this research 21000 images from Kaggle as the dataset are used. The pre-trained models are used for the improvement of recognition accuracy, even with limited training data. Experimental results demonstrate the effectiveness of CNNbased facial recognition models in achieving high accuracy and robustness across varying conditions. This research segmented the process into three parts: Testing, training, and validation. Firstly, the proposed CNN model was trained with this dataset. Moreover, some pre-trained models are also run. They are: Inceptionv3, EfficientNet B0, EfficientNet B6, Xception, and Resnet50. This contributes to the field of facial recognition by offering a comprehensive exploration of CNN-based techniques and addressing real-world challenges.
Application of Modern Supply Chain Management Tools to Power Plants of Bangladesh: A Case Study
(Department of Mechanical and Production Engineering (MPE),Islamic University of Technology(IUT), Board Bazar, Gazipur, Bangladesh, 2017-11-15) Mahmud, Asif; Hossain, Md. Moinul
Power generation sector of a country plays a pivotal role in her overall development and progress. In modern world, as days goes by people are getting depended on technology even more. Bangladesh is the ninth most populous country in the world. In order to turn this overpopulation from a burden to a blessing we need to equip them with necessary technology. To provide technology for this huge amount of people we need a lot of electric power. If we take a look at the most developed countries in the world like USA, CHINA, JAPAN etc. they are also the country which are leading in electricity production. It clearly proves how much impact energy production has on any country’s development. This is where Bangladesh lacks the most. Our country currently ranks 46th in overall electricity production whereas its population ranking is 9th in the whole world. The first idea that comes to our mind in order to increase our country’s overall electricity production is to establish new power plants. But it is a very hazardous, costly, environmentally polluted process. So, it’s clearly not an easy process. There is another thing we can do to increase the electricity production. Electricity production in a power plant is a very complicated and lengthy process integrated with a lot of delicate work to do which keeps a room for improvement in the overall management sector. Modern supply chain tools are currently being used in pretty much every managing sector due to its radical effect. Utilizing it in our power sector can also help in increasing the overall efficiency. It is a relatively new concept in our country. But there are countries around the world using modern supply chain in their power plants and getting positive feedback. Our thesis is a case study about the prospect and goals that can be achieved if modern supply chain tools are applied in power plants of Bangladesh.
Deep Learning Models To Diagnose Brain Stroke
(Daffodil International University, 23-02-18) Mahmud, Asif; Islam, Raisul
Brain Hemorrhage has gotten to be an extreme issue in the world. Brain Hemorrhage is also called ‘Brain Stroke’ in our country. The amount of the stroke may be a particularly difficult process among all of them; the brain stroke is the toughest one. There are three types of strokes. Hemorrhagic Stroke, Ischemic Stroke and Transient Ischemic Attack (TIA).One of the discrete methods that emerged as a first-line, radiation-free method of diagnosing brain stroke is magnetic resonance imaging (MRI). Most of the people (87%) are affected by Ischemic stroke. To distinguish this hemorrhage, we got to do a lab test at therapeutic. But that's exceptionally costly for our country. So, we choose to do something for them. After research, we developed a project which can detect this type of hemorrhage using deep learning neural network based algorithm. When it comes to picture identification, deep learning has made significant progress. To do that we collected raw data (CT Scan Copy) from a number of hospital, we preprocessed it with a help of radiologist and finally trained our model to nail our goal which is detecting hemorrhage stroke. Right now, we have been found that VGG-16 achieves a high rate of accuracy with a minimal level of complexity. It’s about 93 up percentage accuracy in total.Keywords: Brain Stoke, MRI, VGG-16.
Design and implementation of a smart forest environment monitoring system
(BRAC University, 2022-09) Mahmud, Asif; Ahmmed, Tanvir; Prova, Nabila Dohid; Islam, Md Taiubul; Saha, Rony Kumer; Mohsin, Abu S.M.; Rahman, Md. Mosaddequr
Understanding the forestry health and environment—air and water quality, and the density of trees—is the primary objective of our project. From the numerous ways of analyzing the forest, our aim was to find the most effective, but accurate method—the use of a UAV (Unmanned Aerial Vehicle)—to monitor the forest. Given the size, accessibility, and unpredictability of a forest, the UAV has made it possible for us to precisely and efficiently gather the necessary data. Additionally, the health of a forest can be determined by a number of factors—but our project focuses primarily on the three: Air Quality, Water Quality, and the forest's canopy. These factors help us determine the rate of deforestation and the gradual environmental deterioration of a specific forest—over a period of time. In terms of the project's goals, a device was made with sensors for data collection, later the data was analyzed with Standard Data for a comparative idea; and a machine learning model was created to predict the future condition of the forest. Along with this, a camera was used in the device to periodically collect images for processing and calculating the area of the forest. We received fairly accurate test results for our project, and these findings let us comprehend the general state of the forest according to our criteria.
Ensuring Equity In Education: Enhancing Accessibility Of E- Learning Platforms For Targeted Users
(Daffodil International University, 2025-01-18) Mahmud, Asif
In This thesis will find the effectiveness of existing E-Learning platforms, mainly focused on underprivileged learners and the challenges and barriers that low resource learners face in accessing online education. This thesis will identify two main objectives: - The effectiveness of current accessibility features, such as mobile optimization and offline access, and proposing adaptive solutions that enhance elearning accessibility for learners in underprivileged areas. In our study, 500 participant from urban regions involved a quantitative research method to collect data successfully. Suggesting that user friendly mobile designs were highly beneficial by participants. However offline accessibility and the implementation of multilingual support faced challenges. Drawbacks such as internet connections, outdated devices, and limited digital skills were found to break in involvement with online learning. The result shows the importance of having accessibility options. In particular when it comes to offline use and designs that are suitable, for low bandwidth. Additionally, the study suggests incorporating literacy initiatives into online educational programs to assist learners in overcoming challenges. Overall, the study suggests creating fairer e-learning environments that meet the needs of all learners
Real time classroom attendance management system
(BRAC University, 4/18/2017) Islam, Md. Shafiqul; Mahmud, Asif; Papeya, Azmina Akter; Onny, Irin Sultana; Uddin, Jia
Face recognition is a pattern recognition technique and one of the most important biometrics; it is used in a broad spectrum of applications. Classroom attendance management system is one of the applications. Traditional attendance system: roll calling, card punching, paper-based attendance are manual process. It takes a lot of time. To remove hectic of traditional process Real time attendance management system is a better solution. Without physical interaction of human being it gives the attendance of present student in the class. Using Kinect camera we took the video input of the classroom. Detection of human face from the video stream is done by Viola- Jones algorithm. For recognition purpose we tested Speeded Up Robust Features (SURF), Histogram of Oriented Gradients (HOG), Linear Binary Pattern (LBP) feature extraction algorithm and do some comparison between those algorithm for our created dataset. In order to normalization we used Kernel Based Filtering method. In our work, when a face of a student matches with the face of dataset it marked the student as present.