API, SDK or API-SDK Hybrid Implementation

FaceX offers multiple solutions that can be integrated and used to create any kind of facial recognition application. But depending on your use case and project needs you would need to go with our API, SDK or an API-SDK hybrid implementation.

FaceX API

Here, the Client media are analyzed online using FaceX APIs. This is ideal for small scale applications on low budgets.

The current line of FaceX APIs is online and Image-based APIs which can be used for a variety of applications. With Online-Image based APIs,  the application from your end should ensure that a picture containing a face is captured and is posted via an online-based HTTP Post request to the FaceX server.

The use-cases of the API are:

  • Detect & count the number of faces in an Image (Face Detect)
  • Determine approximate Age and Gender of each person (Face Detect)
  • Locate Landmark points of a face (Face Detect)
  • Determine the match confidence between two faces (Face Compare)
  • Convert a face into a Face Template (Face Vector)
  • Create an online image database on FaceX server (Face Search)
  • Search your online database on FaceX server (Face Search)
FaceX SDK

Here, the system works offline at the client’s side and all computation, data storage, face detection, and Face recognition happens at the client-side. This is ideal when the client wants a solution that works offline and when the client requires a faster Face recognition application.

FaceX is running Beta version offline SDKs by way of docker instances, compatible with Linux- ubuntu, which can be used to perform facial recognition from the client-side server. The docker instance comes with a face detector for video-streaming and therefore, the docker instance can be used to perform facial recognition of live-streaming video feeds. The Docker instance can be customized to perform facial recognition of images as well. With Docker instances for Video streaming, the performance of the system is enhanced by using GPU.

Please log-in to the FaceX portal to request beta-access to the FaceX on-prem docker.

API- SDK Hybrid Implementation

In an SDK-API hybrid, part of the application is performed by way of SDKs installed at the client-end and images processed by the SDK are posted to the FaceX API to perform Recognition or other features. There are different API-SDK hybrid implementations offered currently

Offline Active Liveness Detection SDK with FaceX API

Our offline Spoof detection SDK (Liveness Detection by Motion) works on iOS and Android devices and determines if the person is real or not by analysing a video stream. Once the person is determined to be real, a frame from the video feed can be captured and used for Facial Recognition by way of FaceX API

Online Passive Liveness Detection SDK with FaceX API

Our online Spoof detection SDK (Liveness Detection by bright-light modelling) works on iOS and Android devices, where the SDK captures a picture from the device and the picture is posted to the FaceX server for analysis of Liveness. Once the person is determined to be real, the captured picture can be used for Facial Recognition by way of FaceX API

Docker Instance SDK and FaceX API

If you are looking for a more viable implementation of Facial Recognition for Live-streaming videos. A docker instance can be customized to perform only facial detection from the video feed and the detected faces can be posted to the FaceX server for Facial Recognition.

 

Related Articles
Post a Comment