Face Rekognition challenge

Question

Hi ,
During user registration , Our ICEcard app platform provides three face capture 1. Live face with different angles of face, 2&3 user can add their any picture from library . However when we try to scan face to know a registered users live face scan or pictures , the success rate is very less . If same picture as added in app is scanned , then more success rate. Pls guide. 
link of ICEcard app as below from play store.
https://play.google.com/store/apps/details?id=com.rannlab.ice_card.ice_card&pcampaignid=web_share

Ashish_P · Answer

You can refer [Improve accuracy of Amazon Rekognition Face Search with user vectors](https://aws.amazon.com/blogs/machine-learning/improve-accuracy-of-amazon-rekognition-face-search-with-user-vectors/).

You can also check tutorial for [Using Amazon Rekognition for Identity Verification](https://docs.aws.amazon.com/rekognition/latest/dg/identity-verification-tutorial.html) and verify your implementation.

Andrew_M · Answer

There are some [best practices for images used for face oeprations ](https://docs.aws.amazon.com/rekognition/latest/dg/recommendations-facial-input-images.html) and a helpful [diagram for camera position](https://docs.aws.amazon.com/rekognition/latest/dg/recommendations-camera-image-video.html). In summary, plan to use images taken fairly centrally and showing more of the face clearly, as these give better results compared to other types of images. The best practices also give guidance on lighting, image size and resolution, and backgrounds

Face Rekognition challenge

Add your answer

Relevant content