How can I change the cartoon character?

[CheungBH]

Hello. Thanks for your work. I successfully ran your demo webcam.sh, which generates a cartoon character like this. I wonder how to generate such a character, and can I change it? Any there any references about it?

[Arthur151]

Here is the guidance https://github.com/Arthur151/ROMP/blob/master/docs/config_guide.md#mesh_cloth-str By setting the mesh color, you can change the color of naked mesh or change the cloth/appearance of the mesh. We provide a wordrobe for you to choose.

[CheungBH]

Thank you for your reply. I saw the constants. And I wonder where are the parameters controlling the current color?

[Arthur151]

Set it at https://github.com/Arthur151/ROMP/blob/653a0c9de13c7e242bc304147ae6559d1c6ff283/configs/webcam.yml#L36

It incluence the code at https://github.com/Arthur151/ROMP/blob/653a0c9de13c7e242bc304147ae6559d1c6ff283/romp/lib/visualization/create_meshes.py#L51

[CheungBH]

Thank you. I have successfully change the character clothes. By the way, I wonder how to generate the .jpg files in the model_data/wardrobe? Are there any tools you are using?

[Arthur151]

Those are texture UV maps of SMPL mesh. Some of those are obtained from https://github.com/bharat-b7/MultiGarmentNetwork#dress-smpl-body-model-with-our-digital-wardrobe

[CheungBH]

Thank you so much! I ran your demo and it works. However, I can only see one person even I am inputting a multiple people video. Are there any limitations on showing multiple people? My target is that I want to read different video sources, and then merge the extracted human mesh into one image. How do you think of its feasibility?

[Arthur151]

Please set the config setting

show_largest: False

Besides, I think it is highly feasible.

[CheungBH]

Thank you for your help! Maybe I have achieved. BTW, I found that the inference is two times lower when feeding 2 videos, which results from inferring with the function forwarding single image twice. https://github.com/Arthur151/ROMP/blob/653a0c9de13c7e242bc304147ae6559d1c6ff283/romp/predict/base_predictor.py#L43 In this situation, I want to achieve inference with a batch of multiple images and obtain multiple outputs instead. Are there any functions provided? I think it may be contained in the train or validation functions?