By using AWS re:Post, you agree to the Terms of Use

Rekognition bound boxes

0

Hi everyone!

Currently, I've accomplished detecting all the faces from a collection and then generating sub-galleries of each individual subject with all their photos associated with the ruby SDK '~> 1.65' In order to do this, I've indexed the faces of all photos within a collection, list all the faces (https://docs.aws.amazon.com/sdk-for-ruby/v3/api/Aws/Rekognition/Client.html#list_faces-instance_method), then grabbing each face_id recognized and search the faces related to that face_id (https://docs.aws.amazon.com/sdk-for-ruby/v3/api/Aws/Rekognition/Client.html#search_faces-instance_method), and delete the face id used to do the API call and all the returned ones to tell when a new detected subject starts and ends.

My issue is that some indexed faces are returning a wrong bound box not fitting the face properly with a few images, sometimes not even close to the face (I've seen a brick as a bound box). This is weird because internally those "wrong faces" seem to be associated properly with other faces as if the service were just returning wrong box values but doing all the AI jobs properly.

Is there a way to prevent or tweak the call params to prevent this? or is it normal for the API to return sometimes wrong values due to the multiple decimals?

Thanks in advance!

3 Answers
0

Hi, thanks for your reply!

Yes, I forgot to mention that I'm using the formula from the docs https://docs.aws.amazon.com/rekognition/latest/dg/images-displaying-bounding-boxes.html to retrieve the exact box and it's still wrong in some scenarios, displaying bricks or an elbow even.

answered 4 months ago
  • Side note: for further comments to the answer I had sent, you can post as comments, in order not to confuse the voting system and help the community find the most appropriate answers in the future

  • Now, regarding what you said about following the documentation, out of curiosity, if you go the Amazon Rekognition page at AWS Console, then the 'Facial analysis' option at the menu at the left and upload a picture, are you still getting odd results ?

0

Hello, each bounding box width/left and height/top returned by the API is a ratio of the overall width and height, respectively

answered 4 months ago
  • Now, regarding what you said about following the documentation, out of curiosity, if you go the Amazon Rekognition page at AWS Console, then the 'Facial analysis' option at the menu at the left and upload a picture, are you still getting odd results ?

0

Hi, we are happy to work with you to investigate this issue. Can you open a case using this link and share the case ID here. We can then engage directly to troubleshoot.

answered 4 months ago

You are not logged in. Log in to post an answer.

A good answer clearly answers the question and provides constructive feedback and encourages professional growth in the question asker.

Guidelines for Answering Questions