Inference question about myself images

[favorxin]

Hello! I am very interested in your work. I run the demo.py successfully and I had the question about the inputs after read your code in demo.py. The input img is generated by the bbox which is the person's bounding box. I inference the following pic and got a confused result as following(Which bbox is set as the following code). Can I inference the img without the bbox? bbox = [0, 0, original_img_width, original_img_height]

The left image is the original input image and the right image is the rendered_mesh_lixel image.

[mks0601]

Hi, you should use a bbox :(