Deepstream Taking Batches of Images from Minio as an Input

Accelerated Computing Intelligent Video Analytics DeepStream SDK
1

• Hardware Platform: GPU
• DeepStream Version: 5.1-21.02-triton (Docker Container)
• TensorRT Version: 7.0.0-1
• NVIDIA GPU Driver Version: 465.19.01
• Issue Type: questions

Hello, I would like to know how the configuration should look like if I want the Deepstream app to handle batches of images as an input with a pre-determined frequency. The raw images would be updated on a Minio server at a certain time, and the processed images would be stored on the Minio server as well.

Previous Exploration
I know that Deepstream could handle image streams as inputs [In deepstream-app, how to use image stream instead of video stream for detection? - #3 by 1456873666], and I have tried the test-4 app with Kafka. I am not sure if the test-5 app is handy in this case, as it is using RTSP as the input channel [DeepStream Reference Application - deepstream-test5 app — DeepStream 6.1.1 Release documentation]. RTSP is something that I really would like to avoid, yet I could take it if it is the only option.

Any help/suggestion would be appreciated.

3

deepstream-test5 also use the same framework with deepstream-app, it can handle mjpeg input too.

The source code is open source, please check the source code and README inside.

We don’t have module to support Minio server read and write.

4

Thank you!