Face-api.js

About Face-api.js

Face-api.js is a JavaScript library for face detection and face recognition that runs in both the browser and Node.js, built on top of tensorflow.js core. It provides face landmark detection, face expression recognition, and age estimation and gender recognition, exposed through a set of neural network models (such as SsdMobilenetv1, TinyFaceDetector, TinyYolov2, FaceLandmark68Net, FaceRecognitionNet, AgeGenderNet, and FaceExpressionNet) accessible via a high-level API.

Models are loaded by providing a manifest.json file along with corresponding weight shard files, either from a URI in browser environments or directly from disk in Node.js. For Node.js use, the library requires polyfills for browser-specific objects like HTMLCanvasElement, HTMLImageElement, and ImageData, typically provided via the node-canvas package, with optional integration of @tensorflow/tfjs-node for faster native performance.

The library is intended for developers building applications that need face detection and recognition functionality in JavaScript or TypeScript, distributed via npm and usable through a script include in the browser or as an npm package in Node.js projects. It is released under the MIT license.

Key features

• Face detection with multiple detector options (SSD Mobilenet v1, Tiny Face Detector, MTCNN, Tiny YOLOv2)
• 68-point face landmark detection
• Face recognition for identifying and matching faces
• Face expression recognition
• Age estimation and gender recognition

Frequently asked questions