Computer Vision Projects

This workspace contains a collection of computer vision projects, experiments, assignments, and utilities. It is intended as a central hub for research, learning, and reproducible demos related to image and video analysis.

Workspace Structure (high-level)

• malaria_detection_model.ipynb — End-to-end notebook for malaria cell classification.
• assignments/ — Course assignments and supporting notebooks.
• Traffic-Analysis/ — YOLO-based traffic analysis scripts and example video processing.

Example Projects

Below are short, consistent guides for each project. Each entry answers: what the project does, which skills and technologies were used, and what concrete problem or issue was solved.

Malaria detection

• Path: malaria_detection_model.ipynb
• Description: End-to-end notebook that preprocesses blood-smear images, trains a CNN to classify infected vs. healthy cells, and evaluates model performance with metrics and visualizations.
• Skills & Technologies: Image preprocessing, data augmentation, TensorFlow / Keras (or PyTorch), model evaluation, Jupyter notebooks, matplotlib/seaborn for visualization.
• Issues resolved: Created a reproducible pipeline for small medical-image datasets, handled class imbalance with augmentation and weighted loss, improved input normalization to stabilize training.
• Key files: malaria_detection_model.ipynb (notebook), any data/ sample pointers (if present).
• How to run: Open the notebook and run cells top-to-bottom; ensure required Python packages are installed.

Assignments (course exercises)

• Path: assignments/
• Description: Collection of course assignments and example solutions demonstrating core CV techniques and small experiments.
• Skills & Technologies: Fundamental image processing (OpenCV), simple ML model training, data visualization, notebook-based explanations.
• Issues resolved: Step-by-step exercises that clarify image transformations, feature extraction basics, and small classification/regression tasks.
• Key files: assignments/assignment-1/assignment.ipynb and other notebooks under assignments/.
• How to run: Open the relevant assignment notebook and follow the instructions in the first cells.

Traffic Analysis (YOLO)

• Path: Traffic-Analysis/
• Description: Scripts for running YOLO-based detection on videos, extracting object counts, and basic tracking for traffic analysis demos.
• Skills & Technologies: Object detection (YOLO), OpenCV video I/O, simple tracking heuristics, Python scripting for batch processing.
• Issues resolved: Demonstrated applying a pre-trained detector to videos, extracted vehicle counts and basic movement statistics, provided a lightweight pipeline for demoing object detection on recorded footage.
• Key files: Traffic-Analysis/yolo.py, Traffic-Analysis/yolo_video.py.
• How to run: Edit config paths in the scripts and run with python Traffic-Analysis/yolo_video.py.

Getting Started

1. Clone the repository and create a Python environment (recommended: conda or venv).

python -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt  # create this file per-project if needed

2. Open notebooks in JupyterLab or VS Code and follow the instructions at the top of each notebook.

Environment Recommendations

• Python 3.8+ (3.9/3.10 recommended)
• Use conda for easy dependency management where GPU libraries are required.
• Install common CV libraries: opencv-python, scikit-image, tensorflow / torch as needed.

How to Add a New Project

1. Create a new folder at the workspace root with a short, descriptive name.
2. Add a README.md inside the project folder describing purpose, data sources, and how to run experiments.
3. Include a requirements.txt or environment.yml that lists exact dependencies.
4. Add a short entry to this top-level README linking to the project folder, following the template below.

Project entry template (copy/paste)

### Project Name
- Path: [relative/path](relative/path)
- Description: One-sentence summary of the project.
- Skills & Technologies: list of key skills, libraries, and frameworks used.
- Issues resolved: short bullet(s) explaining concrete problems solved or research questions addressed.
- Key files: list of important files to inspect.
- How to run: one-line run command or instruction.

Reproducibility Notes

• Do not commit large datasets. Instead add a small sample, or provide download instructions and checksums.
• Where applicable, pin dependency versions and include seed information for experiments.

Contributing

• Fork the repo or open a branch, add your project folder, and submit a pull request.
• Keep project READMEs focused: objectives, setup, usage, results, and license.

License & Contact

• Each project may include its own license file; the workspace-level README is for orientation only.
• For questions or contributions, open an issue or contact the repository owner.

---