Computer Vision Projects (Python + MediaPipe)

This repository is a collection of real-time computer vision mini-projects built with Python, OpenCV, and MediaPipe.

The goal is simple: detect hands, faces, and body pose from a webcam, then use that data for practical demos like finger counting, volume control, and sign-language action recognition.

What This Project Includes

• Hand tracking and hand landmark detection
• Face detection and face mesh tracking
• Pose estimation
• Finger counting demo
• Hand gesture volume control demo
• Sign-language/action recognition training and inference files

Tech Stack

• Python 3.11
• OpenCV
• MediaPipe
• NumPy
• TensorFlow
• scikit-learn

Project Layout

• HandTrackingModule.py, FaceDetectionModule.py, FaceMeshModule.py, PoseModule.py: reusable detection modules
• *Project.py and *Basics.py files: runnable demo scripts
• AiTrainerProject.py and Sign-Language.py: model training and sign-language workflow
• MP_DATA/: saved training sequence data
• images/ and videos/: assets used by demos

Quick Start

1. Clone the repo
2. Install dependencies
3. Run any demo script

pip install -r requirements.txt
python HandTrackingProject.py

Other examples:

python FaceDetectionBasics.py
python FaceMeshBasics.py
python PoseProject.py
python FingerCountingProject.py
python VolumeHandControl.py

Or use UV

uv sync
uv run HandTrackingProject.py

Same for Other examples

Notes

• Most scripts use your webcam.
• If one script does not run on your machine immediately, check webcam permissions and Python package versions first.

Status

Active learning/project repository with multiple independent experiments.