Computer Vision Researcher | Gait Analysis & Biomechanics
Transforming human motion videos into quantifiable gait metrics, event detections (heel-strike, toe-off, foot-flat), and 8-phase segmentations with full quality control and annotated outputs.
π Research Collaborator β Liverpool John Moores University (LJMU)
Investigating AI-driven human motion analysis for healthcare, safety, and ageing research, blending expertise in computer vision, signal processing, and generative AI.
π§ [email protected]β π LinkedIn
| Project | Description | Tech Stack | Status |
|---|---|---|---|
| Gait Event & Subphase Pipeline | Detects gait events (HS/TO/FF) and segments into 8 sub-phases; exports annotated videos, Excel reports, and QC visualizations. | Python Β· OpenCV Β· SciPy | π§ͺ Internal testing β Public release planned |
| Pose Cleaning Toolkit | End-to-end preprocessing of pose keypoints β gap filling, outlier removal (Hampel), smoothing (SavitzkyβGolay), and anatomical normalization. | NumPy Β· SciPy Β· Pandas | βοΈ Packaging β Docs & examples in progress |
| Gait QC Dashboard | Streamlit dashboard for frame-level gait inspection, event validation, and multi-subject visualization. | Streamlit Β· Plotly Β· OpenCV | π§© Prototype β UI/UX refinement |
π‘ Interested in private demos, data collaboration, or reproducible workflow templates?
Reach out for early access and research partnerships.
- Complete pipelines: video β OpenPose β event detection β segmentation β report generation
- Data cleaning, interpolation, and biomechanical signal normalization
- Automated QC & reproducibility using modular analysis scripts
- LJMU Collaboration: precision gait phase segmentation & transition detection using temporal computer vision
- Signal-based AI: Echo-State Networks, time-series modeling, and hybrid biomechanical features
- Responsible AI: reproducibility, bias assessment, and ethical deployment in healthcare
| Certificate | Platform | Date |
|---|---|---|
| Introduction to Generative AI | Google Cloud | Nov 2024 |
| Vector Search & Embeddings | Google Cloud | Nov 2024 |
| MLOps for Generative AI | Google Cloud | Sep 2024 |
- Human Motion Analysis β Gait biomechanics Β· Spatiotemporal modeling Β· Phase identification
- AI for Healthcare β Clinical motion datasets Β· Rehabilitation Β· Injury prevention
- Generative & Responsible AI β Bias mitigation Β· Reproducibility Β· Biomedical validation
- Mehmood, R., Topham, L., & Khan, W. (in preparation).
Gait Phase Identification and Feature Extraction using Multimodal Computer Vision Approaches.
π Projects Β· π§ Expertise Β· π§° Tech Stack Β· π Research
π¬ Open to collaborations in computer vision for gait, human motion, and digital health.
Letβs build reproducible, data-driven motion analysis systems that actually help people walk better.