MediaPipe
MediaPipe is an AI tool in the Coding & Development category. Useful for computer vision, hand tracking, on-device ml, google, open source.
Google's open source, on-device machine learning toolkit for live and streaming media: hands, faces, poses and more, running right in the browser.
Kit aberto do Google de aprendizado de máquina no próprio aparelho, pra mídia ao vivo: mão, rosto, pose e mais, rodando no navegador.
AITag Review
MediaPipe is Google's toolkit for on-device machine learning on live media. Its Hand Landmarker first finds the palm in the frame, then marks 21 landmarks, the reference points of the hand from the wrist to each fingertip; in video it simply follows the hand from frame to frame and only searches again when it loses it. Everything runs in the browser, so the camera image never leaves the computer. On Sapiens it powers the Particle Control Lab: an open hand pulls the particles into a circle and a peace sign makes them spell SAPIENS, a rule the Lab reads from those 21 points. Measured on our test photos, it marks the 21 points in about 23 ms. Apache 2.0, npm i @mediapipe/tasks-vision.
Features
- Category: Coding & Development
- Tags:computer visionhand trackingon-device mlgoogleopen source
- Website:Visit
- Curated: No
Related tools
An advanced Google platform for developers to build the next generation of applications powered by generative AI and Gemini models. It provides methodologies and tools to integrate multi-modal workflows and intelligent agents into modern software.
REA is an open-source AI agent framework designed to reverse engineer software at every level, from high-level app behavior down to native binaries. It automates complex workflows involving decompilation, behavioral inspection, and binary analysis using intelligent agents. Perfect for security researchers and developers looking to dissect and understand unfamiliar systems with ease.
A high-performance infrastructure platform designed to accelerate AI-driven image, video, and audio generation. It empowers developers to build AI features and creative workflows with industry-leading inference speeds and scalability.
A powerful command-line interface that integrates Google's Gemini AI directly into your terminal to accelerate code development, debugging, and deployment. It enables developers to automate complex tasks and troubleshoot errors using natural language without leaving their workspace.