How to write frames extracted from opencv videocapture to kinesis video stream

Question

I am reading kinesis video stream and performing aws face recognition. However I have to perform tasks like creating rectangle and other post-processing tasks by OpenCV after AWS face reCognition. So, I captured kinesis video with Boto3 library and give input to OpenCV library and performed my operation. Now I am not sure how to write the frame back to kinesis after performing the operation.

This is the reference link showing how to read data from Kinesis video stream. Instead of imshow I need to write back frames to kinesis video stream. How do I do it?

user_11804423 · Answer

Other alternative would be to render the bounding boxes with the video and stream it back to another kinesis video stream using the stream producer.

user_2553321 · Answer

There is a published example to read Kinesis Video data, retrieve AWS Rekognition json output, draw bounding boxes on top of rekognized faces and render it locally or create a new derivative kinesis video stream. Unfortunately it's in java though.

https://github.com/aws/amazon-kinesis-video-streams-parser-library#kinesisvideo---rekognition-examples

How to write frames extracted from opencv videocapture to kinesis video stream

2 Answers