Node.js & practical computer vision recipes


Object detection (car/face):

node-opencv has a number of bundled configurations for feature extraction

var CASCADES = {
   FACE_CASCADE: 'haarcascade_frontalface_alt.xml'
 , EYE_CASCADE: 'haarcascade_eye.xml'
 , EYEGLASSES_CASCADE: 'haarcascade_eye_tree_eyeglasses.xml'
 , FULLBODY_CASCADE: 'haarcascade_fullbody.xml'
 , CAR_SIDE_CASCADE: 'hogcascade_cars_sideview.xml'

Google Cloud vision

First, read “quick start”. Go to API console and enable cloud vision API

Example using official npm package

var gcloud = require('gcloud');

// Authorizing on a per-API-basis. You don't need to do this if you auth on a
// global basis (see Authorization section above).

var vision ={
  projectId: 'vernal-acrobat-120002',
  keyFilename: 'key.json'

//vision.detectText('./image.png', function(err, text) {
//vision.detect('./image2.jpg', ['label', 'face'],

function detect() {
  // this is my phone IP camera
  // though undocumented, you can pass image url
  // and client would request image for you locally
  // first before sending to CV
  vision.detect('', ['label', 'face'], function(err, detections, resp) {
    // we'll do it in a loop in this exemple
    // don't forget to close, otherwise

See more examples on tutorial page