Issue with Detect Faces Activity

Help Marketplace
, , ,
1

We are currently using “Detect faces” activity from UiPath.MicrosoftFace.Activities package and recently microsoft has updated restrictions on the parameters provided as part of the API such as faceid which is set to true by default but we would like to set it to false.

Detect faces activity does not allow the user to customise the parameters, therefore, I would like to use the HTTP web api activity to integrate directly with the API using octet-stream input i.e., binary data/base64 format. I’m receiving a 400 Invalid request error whenever I try to integrate with the API directly. I would like to understand how internal labs (Microsoft Facial Recognition - RPA Component | UiPath Marketplace) is passing the image to the following endpoint as they have developed these packages:

https://australiaeast.api.cognitive.microsoft.com/face/v1.0/detect?returnFaceId=false&returnFaceLandmarks=false&recognitionModel=recognition_04&returnRecognitionModel=false&detectionModel=detection_03&overload=stream

What type encoding or activity or custom code is being used to convert image into the format that is acceptable to the API. This is a critical issue as it came up last week and we had to pause the bot in production until a fix can be applied.

Any help would be greatly appreciated.

Cheers,
Prasanna.