retinaface wasm

Retinaface-WASM

This is a WASM implementation of the Retinaface face detection algorithm. It is based on the Retinaface paper.

This module supports WeChat applet and does not rely on any third-party libraries and can be used directly! (You can manually compress the wasm file with brotli and copy the index.ts code)

Screenshots

Usage

Installation

npm install retinaface-wasm

Example

 import Retinaface , { getWasmFile , env } from 'retinaface-wasm'
import imagePath from './R.jpg'

const wasm = await WebAssembly . instantiateStreaming ( fetch ( 'https://cdn.jsdelivr.net/npm/retinaface-wasm/wasm/' + getWasmFile ( ) ) , env )
const retinaface = new Retinaface ( wasm . instance )

const image = new Image ( )
image . src = imagePath
await new Promise ( ( resolve , reject ) => {
  image . onload = resolve
  image . onerror = reject
} )

const [ data , scale ] = retinaface . processImage ( image )
const result = retinaface . detect ( data , scale )

console . log ( result )

retinaface . close ( ) 

Brotli Compress WASM

node -e " fs.writeFileSync('retinaface-basic.wasm.br', zlib.brotliCompressSync(fs.readFileSync('retinaface-basic.wasm'))) " 

Examples

WeChat Mini Program

WeChat Mini Program Example

478 Landmarks

478 Landmarks Example

Landmark Indexes

Author

Shirasawa

License

MIT