Vision API

API Service that Recognizes and Analyzes Information from Images

Vision API uses the AI image analysis solution of Samsung SDS to analyze and
extract various information from images. With Face Recognition API among
the first to be launched, more APIs are to be provided in the future. Face Recognition API utilizes face recognition technology to detect the location of human faces in images and information on face coordinates and facial features. Similarity measure between two images is also supported.

Overview

  • Globally Acclaimed Facial Recognition Technology

    The facial recognition technology of Samsung SDS AI image analysis solution ranked 6th globally and 1st domestically from the Face Recognition Vendor Test(FRVT) and
    7th globally and 1st domestically for achieving high accuracy from the mask effect test, both of which conducted by National Institute of Standards and Technology(NIST). It continues to maintain top positions on the global stage and the first place within Korea while continuing to train data sets for better accuracy and performance.

  • Face Detection and Comparison

    By recognizing as many as 100 faces, Face Recognition API extract face coordinates and facial features (eyes, nose, and lips) and measure facial similarity by comparing one face with several faces. This service will be the foundation to developing various systems such as the access control system, smart CCTV, and intelligent control system.

  • Convenient Data Integration

    Users can easily integrate facial data extracted from Vision API to the application in use. With Object Storage on the Samsung cloud platform, it is also convenient to upload images to analyze and store analysis results.

01

04

Service Architecture

  • User → Apply for service & Issue OpenAPI key → Vision API
  • AI Utilization System/ML Development Environment → API call/response using a previously-issued key → Vision API
Vision API
  • SCP Console: Manage Authentication Key, Apply for Service, Check Usage, View Usage, Object Storage
  • Vision API Management: API Endpoint, Monitoring/Metering
  • Vision API Engine: Face Detection, Face Comparison

Key Features

  • Supported API

    - Face Recognition API

  • Supported features

    - Face detection : Extract face coordinates and location of eyes, nose and lips
    - Face comparison : Determine facial similarity by comparing the extracted information 1:1 or 1:N

  • Limitation

    - Image format : JPEG, PNG
    - Minimum image size : 80x80 pixel
    - Maximum image size : 4,096x4,096 pixel
    - Maximum capacity
      · Base64 image : 5MB
      · Image stored on Object Stage : 15MB
    - Maximum number of faces detected : 100

  • View API call status

    - View call status by API and by day/week/month
    - View the number of calls by API (All/success/error)
    - View response time for successful API

Pricing

    • Billing
    • Unit price applied based on the number of calls by API
Let’s talk

Whether you’re looking for a specific business solution or just need some questions answered, we’re here to help

Get in Touch
Share