search

muelea / shapy

CVPR 2022 - Official code repository for the paper: Accurate 3D Body Shape Regression using Metric and Semantic Attributes.
https://shapy.is.tue.mpg.de/
305 stars 44 forks source link

Running Attributes to Shape (A2S) on my data #23

Open hongsukchoi opened 1 year ago

hongsukchoi commented 1 year ago

Hi, Thank you for your great work.

I am curious whether I can directly use your pre-trained model to test only my data. For example, I know the person's height and weight, and also the image of him.

Can I directly evaluate A2S model and get SMPL-X betas by changing the commands?

Or do I have to re-trained the model by myself?

Thank you:)

muelea commented 1 year ago

Hi,

thanks for trying out SHAPY.

If you want to use an image and measurements jointly you would have to train a new model. There is a workaround that theoretically could work. Note that I haven't tried it yet. First, you get SMPL-X betas for the image from SHAPY (regressor) and pass them to S2A to get attributes. Then use AHW2S to get refined betas.

I think that might be interesting for others as well. If it works you can create a pull request.

Best, Lea