search

Daniil-Osokin / lightweight-human-pose-estimation-3d-demo.pytorch

Real-time 3D multi-person pose estimation demo in PyTorch. OpenVINO backend can be used for fast inference on CPU.
Apache License 2.0
656 stars 138 forks source link

Interpreting raw model output #40

Closed irwinwilliams closed 3 years ago

irwinwilliams commented 3 years ago

Hey, I've just started working with ML. I'm using ML.net and using a container, converted the model in .pth to .onnx. So far, so good. I've been able to run the model in C#, but I don't know how to convert the results to strings that represent the various poses found. I suspect I should use parse_poses, but I'm new to python, so it's a bit hard for me to parse. What's it doing? Taking the features output and deriving the relevant strings? TIA.

Daniil-Osokin commented 3 years ago

Hi! Does explanation in #28 help?

irwinwilliams commented 3 years ago

I think I'm a step before that. I have the features, heatmaps and pafs. I'm in C#, so trying to convert the code that gets pose information - not trying to overlay on result image tbh. Instead, I'd like to simply report via text on what poses were found.

Daniil-Osokin commented 3 years ago

Looks like this is what you need. It refers to the function which parses network output and returns either 2D coordinates or 3D.

irwinwilliams commented 3 years ago

Hey @Daniil-Osokin did you mean to include something in your last comment?

Daniil-Osokin commented 3 years ago

I mean #28 refers to the function which parses network output and returns either 2D coordinates or 3D.

Daniil-Osokin commented 3 years ago

Hope, it helped.