A computer vision project for detecting study-focus signals from Zoom-like webcam frames and producing interpretable outputs: gaze, headphones, background/privacy, and object in hand.
- Presentations: interim milestones and slides
- Folder:
presentations/
- Folder:
- Notebooks: code used for synthetic data creation, EDA, and baseline evaluation
- Folder:
notebooks/
- Folder:
- Documentation: labeling guidelines and related docs
- Folder:
docs/
- Folder:
- Sample dataset (for review): a small subset of real + synthetic images with labels
- Folder:
data/
- Folder:
notebooks/README.md- Data generation:
notebooks/5_1_data_generation.ipynb - EDA:
notebooks/5_2_eda.ipynb - Baseline evaluation:
notebooks/5_3_baseline_evaluation.ipynb
docs/README.md- Guidelines (PDF):
docs/Labeling_Guidelines.pdf
Work in progress. The interim submission includes:
- synthetic data generation pipeline
- EDA on the labeled dataset
- baseline model training/evaluation and exported metrics/plots