Please provide the following information when requesting support.
• Hardware (T4/V100/Xavier/Nano/etc)
Xavier
• Network Type
2/3D action recognition
• TLT Version (Please run “tlt info --verbose” and share “docker_tag” here)
3.22.05
For this 2/3D action recognition, the engine can detect action of individual person in image with multiple persons?
Primary use case intended for this model is to recognize the action from the sequence of RGB frames and optical flow gray images.
See more in
https://docs.nvidia.com/tao/tao-toolkit/text/action_recognition_net.html#running-inference-on-the-model
https://docs.nvidia.com/tao/tao-toolkit/text/action_recognition_net.html#preparing-the-dataset
No my question is does the model detect action of multiple persons in an image OR only one person in an image?
It detects one person.
The algorithm itself (since using rgb and optical flow in training) supports only single person?
Those activity detection algorithms support multiple persons detection in an image.
Correct my previous comment.
Please see below.
The action recognition model in TAO does not do detection.
It just classify the input sequence.
So if multiple person in one image doing the same action, it might work.
If multiple person in one image are doing multiple action, it can only give one classification and it might be not correct.
Thanks the algorithm is implementation on activity recognition of the scene in image. Not activity detection of individual person in an image.
Yes, the algorithm is not the same as your shared link.
This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.