Face Match API

This API verifies whether two provided images are same or not. It returns a boolean value indicating whether the faces match.

HTTP Method and Endpoint

POST | https://api.videosdk.live/ai/v1/face-verification/verify

Headers Parameter

Authorization

values : YOUR_TOKEN_WITHOUT_ANY_PREFIX

This will be a JWT token generate using VideoSDK ApiKey and Secret.

Note : the token will not include any prefix such as "Basic " or "Bearer ". Just pass a token as value.

You can generate a new token by referring to this Guide: Generate Auth token

Content-Type

values : application/json

This is useful for json body parameters, so that VideoSDK servers can understand that the incoming body parameter will be a JSON string.

Data Parameter

Base64 Encoding Format

The images should be converted from binary format to a Base64 string

Base64 Encoding

NodeJS

• NodeJS
• PHP
• Python

const base64Data = Buffer.from(image).toString('base64');

Images must be Base64 encoded in the following format:

data:image/jpeg;base64,${Base64data}

Request Format

The request body should contain two Base64-encoded images in the following format:

const data = {
  "img1": "data:image/jpeg;base64,${Base64data}", // Base64-encoded image 1
  "img2": "data:image/jpeg;base64,${Base64data}"  // Base64-encoded image 2
}

Sending Image Comparison Request

The request is made using the following code:

Axios POST Request

NodeJS

• NodeJS
• PHP
• Python

const response = await axios.post(url, data, { headers });
console.log(response.data);

This sends the encoded images to the specified URL for comparison.

Code Snippet for API Integration

Below is a code snippet demonstrating how to use the Face Match API with Node.js:

import axios from 'axios';
import { readFileSync } from 'fs';
import path from 'path';
import dotenv from 'dotenv';

// Load environment variables from .env file
dotenv.config();

// Get the directory of the current file
const __dirname = path.resolve();

// Function to read image files and convert them to base64
function getBase64Image(imageName) {
  const filePath = path.join(__dirname, 'image', imageName);
  console.log(`Resolved file path: ${filePath}`);
  try {
    const image = readFileSync(filePath);
    const base64Data = Buffer.from(image).toString('base64');
    return `data:image/jpeg;base64,${base64Data}`;
  } catch (error) {
    console.error(`Error reading image file: ${filePath}`, error.message);
    throw error;
  }
}

// Function to perform Face Match
async function faceMatch(ovdImageName, selfieImageName) {
  const url = 'https://api.videosdk.live/ai/v1/face-verification/verify';
  const headers = {
    'Authorization': `${process.env.API_KEY}`,
    'Content-Type': 'application/json'
  };

  const data = {
    img1: getBase64Image(ovdImageName),
    img2: getBase64Image(selfieImageName)
  };

  try {
    const response = await axios.post(url, data, { headers });
    console.log('Face Match Result:', response.data);
  } catch (error) {
    console.error('Error during face match:', error.response ? error.response.data : error.message);
  }
}

// Main function to run the API calls
async function main() {
  const ovdImageName = 'img1.jpg';
  const selfieImageName = 'img2.jpg';

  await faceMatch(ovdImageName, selfieImageName);
}

// Run the main function
main().catch(error => console.error('Error in main function:', error.message));

Examples of Face Match API Usage

Example 1 : Matching Images of the Same Person

In this scenario, we take two different pictures of the same individual. For instance:

Image 1

Image 2

When these images are sent to the Face Match API, the expected output would be:

Face Match Result: { "verified": true } 

This result indicates that despite differences in lighting, angle, or expression, the API successfully recognizes that both images depict the same individual.

Example 2 : Matching Images of Different Persons

In this case, we compare two photos of different individuals. For example:

Image 1

Image 2

When these images are processed through the Face Match API, the expected output would be:

Face Match Result: { "verified": false }

This result demonstrates the API's ability to differentiate between distinct individuals accurately.