Real time AI hand pose detection and Gesture recognition app build by using ReactJS and react webcam

HandyPoses is a hand pose estimation dataset generator using parametric models for precise domain control, overcoming limitations of manual generation, GANs, and video-game-assisted techniques. The system enhances dataset diversity through variations in arm position, orientation, joint angles, skin tone, background, lighting, and shininess.

Real time AI hand pose estimation and positioning app build by using ReactJS and react webcam

Navigate using hand gestures

An illustrator app using hand pose detection from TensorFlow

Uso de TensorFlow.js para llevar a cabo la predicción de imágenes mediante el uso del modelo MobileNet y la clasificación usando el algoritmo K-Nearest Neighbor y 🖐️PoseDetection. Permite la identificación de objetos en imágenes de fuentes como Picsum, e integra la habilidad de cambiar de manera dinámica streams en tiempo real desde la webcam

A small fullstack project that utilizes a custom Tensorflow model and mediapipe to detect hand gestures which control the cube. The project uses a custom cube modelled in Blender and is displayed on the frontend using React Three Fiber. The project also includes a small backend to serve the Tensorflow.js model to the frontend.