Please vote for this issue by adding a đź‘Ť reaction to the issue. This will help the community and maintainers from the Amazon Chime SDK team to prioritize this request.
Please do not leave "+1" or "me too" comments, they generate extra noise for issue followers and do not help prioritize the request.
Please go through the existing issues to see if your request has already been recommended.
Please do not use this template for bugs or support. Each individual repository has a PR template, documentation and resources for troubleshooting.
If you are interested in working on this issue or have submitted a pull request, please leave a comment.
I know that it should be possible to asynchronously analyze 5-second fragments that are produced by Media Capture Pipelines but I would like to avoid that 5 seconds delay.
Community Note
Tell us about your request
What do you want us to build?
This request is for amazon-chime-sdk-js, amazon-chime-sdk-ios, amazon-chime-sdk-android
Tell us about the problem you are trying to solve and why is it hard?
I would like to synchronically analyze the video of the Chime SDK meeting attendee by Amazon Rekognition(examples: https://github.com/aws-samples/amazon-rekognition-virtual-proctor#architecture, https://github.com/brain-power/Brain-Power-Amazon-Fidgetology#components). Video stored in KVS can be processed by Amazon Rekognition stream processor or Amazon Rekognition Custom Labels(through GetImages endpoint).
I know that it should be possible to asynchronously analyze 5-second fragments that are produced by Media Capture Pipelines but I would like to avoid that 5 seconds delay.
I managed to execute Amazon Rekognition Custom Labels on frames sent from the VideoFrameProcessor(https://github.com/aws/amazon-chime-sdk-js/issues/1731#issuecomment-1120481000) but still, possibility to analyze the video captured by Media Capture Pipelines is required solution.
I'm considering sending video captured like in this example https://github.com/aws-samples/amazon-chime-sdk-recording-demo
to KVS. But I guess it will be hard to implement and maintain on a professional level.
How are you currently solving a problem?
Are there any workarounds or mitigations you have used? Why is it not working for you? N/A
Additional context
Anything else we should know? N/A