View Proposal

Proposer
Mohammad Ammad-Uddin
Title
Design, Implementation and Evaluation of AI-Based Real-Time Human Movement Analysis Using Computer Vision
Goal
To investigate, design, implement and evaluate an AI-based computer-vision system for analysing human movement in real time. The student will identify and select an appropriate application domain and develop a prototype that uses human pose information to recognise, assess or provide feedback on selected movements.
Description
Human pose estimation and computer-vision techniques can be used to analyse body movement from ordinary camera input. These techniques can support a wide range of intelligent applications in areas such as fitness, sport, workplace ergonomics, activity recognition, gesture analysis and movement training. This project will investigate the application of AI-based human pose estimation to a selected real-time movement-analysis problem. The student will first review potential application areas and select an appropriate problem based on practical relevance, availability of suitable data, feasibility of real-time implementation, technical depth and the available project timeframe. The selected application will be analysed to determine which movements, pose features and temporal characteristics are important. The student will investigate and select a suitable pose-estimation approach to extract body landmarks from recorded or live camera input. Appropriate movement-analysis techniques may include geometric rules, joint-angle analysis, machine-learning classifiers, temporal models or hybrid approaches. The student should develop a prototype capable of analysing a selected set of movements relevant to the chosen application. Depending on the selected problem, the system may perform movement recognition, repetition counting, posture assessment, movement-quality classification, anomaly detection or real-time feedback. Public datasets may be used where suitable; a limited project-specific dataset may also be created where feasible and ethically appropriate. The final system should demonstrate real-time or near-real-time analysis using a webcam, smartphone camera or similar input source. The project should emphasise investigation and justification of design choices rather than implementation of a predetermined exercise or fixed model. Expected Project Work • Investigate and compare possible human-movement application areas. • Select and justify a feasible application and a limited set of target movements. • Identify or prepare appropriate data for development and testing. • Select and implement a suitable pose-estimation and movement-analysis pipeline. • Develop a working real-time or near-real-time prototype. • Perform experimental evaluation and critically analyse strengths, limitations and failure cases. Example Applications (Illustrative Only) • Fitness exercise analysis, such as assessing push-ups, squats, lunges or jumping movements. • Sports movement analysis, such as evaluating a golf swing, tennis stroke, football kick or athletic movement. • Workplace or study-posture monitoring using body-pose information. • Human activity recognition from live or recorded video. • Gesture or movement recognition for touch-free interaction. • Dance, yoga or movement-training feedback. • Rehabilitation-style movement monitoring as a non-clinical research prototype, subject to appropriate scope and ethics.
Resources
Background
Url
Difficulty Level
Challenging
Ethical Approval
None
Number Of Students
3
Supervisor
Mohammad Ammad-Uddin
Keywords
artificial intelligence, computer vision, human pose estimation, movement analysis, activity recognition, machine learning, real-time vision, mobile ai
Degrees
Bachelor of Science in Computational Sciences and Software Engineering
Back to List