Rekognition video stream - ability to check eye status

Question

I am trying to use this sample to recognize eye status of an operator 
I managed to connect my camera as source of a kinesis video stream then send data to rekogniton 
I used the person recognition sample to make sure i have the right person and it works I can see the name in the json format 
now I am trying to check eye status (open or close) but command is not working - the most important part is eye status  1) what is wrong in below to just get eye status without even caring who is the person 
2) can I recognize the person and also see the eye status  ? (see the commented out # section for example)

my code:

aws rekognition create-stream-processor \
  --input "KinesisVideoStream={Arn=$VIDEO_ARN}" \
  --name kvs-operator-processor \
  #--settings "FaceSearch={CollectionId=$COLLECTION_ID,FaceMatchThreshold=50.0}" \
  --setting "--settings "{"GeneralLabels":{"LabelInclusionFilters":["person"]}}
 ----attributes "ALL"
  --role-arn $ROLE_ARN \
  --stream-processor-output "KinesisDataStream={Arn=$DATA_ARN}"

Andy Shahbazi · Answer

Hi Chris, 
You can't include attributes in this construct

aws rekognition create-stream-processor \
  --input "KinesisVideoStream={Arn=$VIDEO_ARN}" \
  --name kvs-workshop-processor \
  --settings "FaceSearch={CollectionId=$COLLECTION_ID,FaceMatchThreshold=50.0}" \
 [ --attributes "ALL" \]
  --role-arn $ROLE_ARN \
  --stream-processor-output "KinesisDataStream={Arn=$DATA_ARN}"

I removed attribute and your construct works but wont provide eye  status - do you need to find the person in your collection first then check the eyelids ?

Rekognition video stream - ability to check eye status

回答を追加する

関連するコンテンツ