http://hdl.handle.net/123456789/17505 2026-10-17T02:51:24Z 2026-10-17T02:51:24Z 03-135222-047, Sumavia Ali http://hdl.handle.net/123456789/21741 2026-10-06T05:41:08Z 2026-06-01T00:00:00Z

Human-Computer Interaction Based Usability Study of Mobile Banking Applications 03-135222-047, Sumavia Ali Mobile banking apps are really important to us in our lives especially in countries where more and more people are using mobile phones. Even though we use them all the time people still have a lot of trouble with these apps like it is hard to navigate and figure out how to do things. They do not have enough features to help older people or people with disabilities and the security measures are not clear and that makes people not trust them and not want to use them. We undertook a detailed analysis of the mobile banking applications through the HCI approach that studies interaction between users and computer systems. Multiple approaches have been used for the analysis of the applications, among which include heuristic evaluation that makes use of ten usability heuristics for determining user-friendliness. Other approaches included user testing whereby tasks are completed by the user and questionnaires filled out. Further, the applications underwent testing on various mobile platforms to determine their functionality and suitability. We ensured that there was diversity in our experiment through selection of participants of varied age, gender, culture, and proficiency in the use of technology. While conducting the test, we assessed the ease with which users could accomplish fundamental actions like logging into their accounts, viewing their balance, and making transactions. We also tested the ease with which users could accomplish more complicated activities such as setting up preferences and using the accessibility options available to assist disabled users. In the analysis process, we applied both quantitative and qualitative techniques, thus incorporating numerical information with user reviews. v From our findings, improved mobile banking interfaces were designed using Figma software, and then we proceeded to test these prototypes through user experience. The use of advanced technology such as biometrics (fingerprint and facial recognition), as well as voice commands, was explored as part of the research process. This project will help to guide future designers of mobile banking applications on how to design such apps successfull. Dr Junaid Nasir Qureshi

2026-06-01T00:00:00Z 03-135222-051, UMAMA ZINOOR http://hdl.handle.net/123456789/21743 2026-10-06T05:46:16Z 2026-06-01T00:00:00Z

Deep Learning based Leukemia Detection System using Microscopic Images 03-135222-051, UMAMA ZINOOR Blood cancer creates a disruption to healthy white blood cell production in bone marrow; this type of cancer is known as leukemia. Current methods used to diagnose leukemia rely on manually analyzing blood smears, but have a slow turnaround, require a high level of expertise and suffer from inter-observer variability. These limitations create an urgent need for automated, reliable methods for classifying leukemia sub types (ALL, AML, APML, CLL and CML) using a data set called LeukemiaAttri. This study presents innovative approach based on a hybrid deep learning architecture to assist with the automatic classification of five sub types of leukemia using this previously unexamined data set.. Five CNN models and seven Vision Transformer models were trained across two augmentation phases. MobileNetV2 achieved the best CNN accuracy of 95.58% while Swin Base Patch4 Window7 achieved the best ViT accuracy of 95.02%. The two models were fused using Feature-wise Linear Modulation (FiLM), which allows the ViT branch to intelligently modulate CNN features at three scales through learned parameters, enabling richer feature interaction than simple concatenation. A 97% accuracy on the final test set and an AUC-ROC of 0.9993 were achieved with the fused model. The external validation based on 4,915 independent images demonstrated the model's ability to generalise, having an accuracy of 94.85% and AUC-ROC of 0.9949. Grad-CAM++ heat maps confirmed that the model attends to clinically relevant regions of cells, supporting interpretable and trustworthy predictions.The system was deployed as LeukoDetect AI, a Flask-based web application offering real-time classification, confidence scores, and Grad-CAM++ visualizations to support clinical decision making. Dr Shahid Mehmood

2026-06-01T00:00:00Z 03-135222-015, Muneeb Javed http://hdl.handle.net/123456789/21748 2026-10-06T05:55:29Z 2026-06-01T00:00:00Z

SMART AI-POWERED EXPENSE TRACKING SYSTEM 03-135222-015, Muneeb Javed The Smart AI-powered Expense Tracking System is a modernized personal financial management tool aimed at overcoming limitations associated with classical budget applications. The first main purpose of this innovative application is to get rid of problems connected with excessive reliance on human intervention and outdated ways of processing information. To this end, the Smart AI Expense Tracking System incorporates modern internet-based tools, rule-based financial calculations, and an LLM model of conversation with access to external APIs. The most distinctive characteristic of the proposed system is its artificial intelligence component. The AI-powered financial advisor makes it possible for people to communicate with the software using regular language. It means that any person can use natural language to find out more about his/her expenditure habits, saving practices, budgeting methods, and financial behaviors in general. Based on the provided context, the artificial intelligence processes all queries and provides answers to them. Geolocation features are integrated into the app's functionality to improve its ability to understand context from the point of view of user behavior. Through recognizing the presence of the user at certain places (such as retail shops, restaurants, or transportation points), it becomes possible to deduce the spending activities associated with being in these locations, which helps to ensure that expenses are logged promptly. Therefore, the need for manual data entry is significantly reduced and data accuracy increased. FastAPI serves as the foundation for the backend architecture of the application, ensuring excellent speed and performance capabilities of the server. All functions are performed by this framework, which includes user authentication, spending and budget tracking, requests related to AI algorithms and input validation. As for data storage, a light and effective SQL database named SQLite is used to store user data, including user profiles, budgets, and spending locations. In addition to that, the system is built with a significant focus on data visualization and UX. Data concerning finances of the users is provided by means of interactive dashboards, graphs, and chart representations, which make it easier for the users to get insights into their spending. Thus, this system makes visualizations out of complicated data in order to help users analyze their spending patterns. Mr. Tahir Iqbal

2026-06-01T00:00:00Z 03-135222-034, M. Taha Munir http://hdl.handle.net/123456789/21749 2026-10-06T05:57:09Z 2026-06-01T00:00:00Z

ReMind 03-135222-034, M. Taha Munir The rise of mental health issues among university students and young professionals may stem from academic stress, social isolation, and access to low-cost counselling, which have created greater stress, anxiety and depression symptoms. While there are digital wellness tools available, many are not clinically grounded, have limited capabilities, or require a subscription to increase accessibility for those who could benefit most. This thesis presents the design, development, and testing of ReMind, a cross-platform mobile application built using Flutter and Firebase. The app offers a research-backed approach to stress management and emotional awareness through features such as an emoji-based mood tracker, the clinically validated DASS-21 assessment, guided breathing exercises, the 5-4-3-2-1 grounding technique, interactive mini-games, and a CBT-based resilience training module. It also includes journaling, gratitude practice, an AI chatbot for emotional support with a multi-provider fallback system, and a dashboard for tracking mood and assessment progress. All features were developed with guidance from a certified clinical psychologist to ensure clinical appropriateness. There are three authentication methods mainly, email and password, Google Sign-In and anonymous access with the option to also migrate to the cloud for user privacy. The quality assurance process was conducted over a set of approximately 100 test cases, with a 100% success rate in these cases, ensuring the reliability and functional correctness of the system. This project helps meet UN Sustainable Development Goal 3(Good Health and Well-being) through providing a free, self-guided, mental health support tool. Dr Junaid Nasir

2026-06-01T00:00:00Z