Karim Iskakov
Karim Iskakov
Hi, @Samleo8! You’re correct about heatmaps. As a backbone for CMU we used a model, pretrained on COCO dataset (from here https://github.com/microsoft/human-pose-estimation.pytorch/blob/master/README.md). You still need to evaluate Algebraic model to...
Looking at the images above, I think there can be 3 possible problems: 1. Wrong location of the cube. Maybe human doesn’t fully fit into the cube 2. Something wrong...
Hey, @agenthong! Yes, this tensor consists of 3D locations of joints. You can project it back to image using calibration matrix of your camera.
You can find example [here](https://github.com/karfly/learnable-triangulation-pytorch/blob/master/mvn/utils/vis.py#L138).
@agenthong To convert 3D points to your camera coordinate system, you need to apply rotation (R) and translation (t) to these 3D points.
@chaisheng-dawnlight Try to visualize these 3D points without any processing with plt scatter 3D function (https://www.geeksforgeeks.org/3d-scatter-plotting-in-python-using-matplotlib/)
@agenthong Actually in the coordinates of the 1st camera. It usually equals to world coordinates.
@chaisheng-dawnlight What plot do you get from this code?
Hi, @JonathanLehner. I also noticed such instability during training an Algebraic Model, but I got such error only after >10 epochs.
Hi! Nice idea with docs as context. I'll look at it, cuz now I'm not really familiar with LangChain and other