Pytorch3D: 2D projection object boundary mask and visual quality

[Divya_B (Divya)]

Hi Team,

I am a newbie to pytorch3d exploring supercool notebooks!
I have the following questions on pytorch3d/docs/tutorials/render_textured_meshes.ipynb at main · facebookresearch/pytorch3d · GitHub

1. image_grid(images.cpu().numpy(), rows=4, cols=5, rgb=True) previews the 3D object projection on to 2D view. Is there a generalizable way to extract accurate 2D object boundary mask?

2. img1 = images.cpu().numpy()[0]; plt.imshow(img1) yields image of different visual quality(slightly faded and grainy) compared to the plot from #1 . why is there a difference?

Thanks!