Rekognition API: IndexFaces vs. DetectFaces

Question

Hello,

The 2 API from Rekognition share a same structure at some level of depth in the Response: FaceDetail (the object that contains lots of info about a face: ageRange, boundingBox, sunglasses, mouthOpen, emotions, gender, beard, pose, confidence, landmarks, mustache, smile, quality, eyesOpen, eyeDirection, eyeglasses, faceOccluded).

Could a developer consider that the 2 API have identical structure and values, for same: picture and face in the picture ?

Thank you,  
Mihai ADAM

Accepted Answer

Hi, rePost-User-4717319 / Mihai ADAM

The [DetectionAttributes](https://docs.aws.amazon.com/rekognition/latest/APIReference/API_IndexFaces.html#:~:text=Required%3A%20Yes-,DetectionAttributes,-An%20array%20of) (**IndexFaces**) and [Attributes](https://docs.aws.amazon.com/rekognition/latest/APIReference/API_DetectFaces.html#:~:text=in%20JSON%20format.-,Attributes,-An%20array%20of) (**DetectFaces**) arrays do contain the same value types.

Here are the different API calls for [IndexFaces](https://docs.aws.amazon.com/rekognition/latest/APIReference/API_IndexFaces.html) and [DetectFaces](https://docs.aws.amazon.com/rekognition/latest/APIReference/API_DetectFaces.html) - the names on the left menu are the specific API calls (request/response) you can use for each.

If this answers your question, please choose this as the Accepted Answer so re:Post users may benefit - Thank you

Answer

Hi, rePost-User-4717319 / Mihai ADAM

Regarding your follow up questions, the BoundingBox values provided by [CompareFaces](https://docs.aws.amazon.com/rekognition/latest/APIReference/API_CompareFaces.html) has Face, SourceImageFace & UnmatchedFaces for BoundingBox info here -> https://docs.aws.amazon.com/rekognition/latest/APIReference/API_CompareFaces.html#API_CompareFaces_ResponseSyntax

[IndexFaces](https://docs.aws.amazon.com/rekognition/latest/APIReference/API_IndexFaces.html) has Face, FaceDetail & UnindexedFaces for BoundingBox info here -> https://docs.aws.amazon.com/rekognition/latest/APIReference/API_IndexFaces.html#API_IndexFaces_ResponseSyntax

[DetectFaces](https://docs.aws.amazon.com/rekognition/latest/APIReference/API_DetectFaces.html) has FaceDetails for BoundingBox info here -> https://docs.aws.amazon.com/rekognition/latest/APIReference/API_DetectFaces.html#API_DetectFaces_ResponseSyntax

As a side note, I was curious if you've tried using [ListFaces](https://docs.aws.amazon.com/rekognition/latest/APIReference/API_ListFaces.html) yet.  All AWS APIs typically have a Describe/List/Get read API (or combination of those), for reading values also.  It looks like ListFaces responds with  BoundingBox information as well.

If you continue to see a mismatch in the BoundingBox response output between APIs, I'd recommend opening a support case in the AWS account you're running the API commands in/against.  Here are the instructions on how to open a support case:

To open an AWS Support case, you must either be signed into your AWS account as the root account owner, or have IAM permissions to open a support case.

* Sign into your AWS account and navigate to AWS Support Center (you can search for this in the search bar, at the top of the Account console).
* Choose Create case.
* On the Create case page, choose Rekognition and fill in the required fields on the form.
* After you complete the form, under Contact options, choose either Web for an email response, or for Business/Enterprise support customers, Phone to request a telephone call from an AWS Support representative or Web, for instant chat.

I hope this is helpful.

If this helps answer your question, please choose this as the Accepted Answer so others on re:Post may benefit - Thank you

Best Regards,

AWS-Eric

Rekognition API: IndexFaces vs. DetectFaces

Relevant content