How kpjadhav30/gesture_x works
GestureX is a webcam-based application with two user-facing modes: sign recognition and gesture-driven Windows control. The documented runtime flow turns hand landmarks into model predictions, then presents a character or performs an action. The graph also preserves the separate dataset-preparation and model-training capabilities visible in sampled source. The README documents the overall inference flow; sampled source shows parts of the UI, training, and prediction implementations, but not all wiring between them.
Main entry points: app.py Streamlit app · realtime_prediction.py Sign prediction · gesture_pridiction.py Gesture prediction · landmark_extractor.py Landmark extraction
Components 15
Real-time interaction
- Streamlit app
ui/app.py - Hand tracking
src/hand_tracking/hand_detection.py - Landmark extraction
src/hand_tracking/landmark_extractor.py
Sign recognition
- Sign prediction
src/sign_recognition/realtime_prediction.py - Sign models
models/sign_language_model.py
Gesture control
- Gesture prediction
src/gesture_control/gesture_pridiction.py - Gesture models
models/gesture_control_model.py
Data and training
- Sign preprocessing
src/datasets/dataset_preprocess_sign.py - Gesture model builder
src/gesture_control/build_gesture_model.py - Gesture preprocessing
src/datasets/dataset_preprcess_gesture.py - Sign data
outputs/preprocessed_signs - Gesture data
outputs/gesture_control_dataset.csv
Other components
- User
- Webcam
- Windows input
Connections 18
- User → Streamlit app: selects mode
README.md - Webcam → Streamlit app: streams frames
README.md - Streamlit app → Hand tracking: detects hands
- Hand tracking → Landmark extraction: extracts landmarks
- Landmark extraction → Sign prediction: provides features
- Landmark extraction → Gesture prediction: provides features
- Sign prediction → Sign models: predicts signs
README.md - Sign prediction → Streamlit app: shows character
README.md - Gesture prediction → Streamlit app: shows prediction
ui/app.py - Gesture prediction → Windows input: triggers actions
README.md - User → Gesture model builder: collects gestures
src/gesture_control/build_gesture_model.py - Gesture model builder → Gesture data: writes samples
src/gesture_control/build_gesture_model.py - Gesture data → Gesture preprocessing: prepares features
- Gesture preprocessing → Gesture models: trains models
- Gesture models → Gesture prediction: supplies classifier
- Sign data → Sign preprocessing: prepares features
src/datasets/dataset_preprocess_sign.py - Sign preprocessing → Sign models: trains classifiers
- Sign models → Sign prediction: supplies classifier