Thanks for your great work, I'm reading the paper and the code. I'm confused that why focal length is set so large.
When I visualized the 3D global locations, the z values are times larger than x and y. I calculated global locations like this: pred_joints += pred_cam_t. And I printed out the min and max values of [x, y, z] of one person in demo video, which are: min[x, y, z] = [-2.5283, 0.3676, 134.07 ], max[x,y,z]=[2.5229, 0.9773, 243.8021]. So is it correct that z values times larger than x, y? Or can I just set the focal_length in base.py to a smaller one without retraining?
Thanks for your great work, I'm reading the paper and the code. I'm confused that why focal length is set so large.
When I visualized the 3D global locations, the z values are times larger than x and y. I calculated global locations like this: pred_joints += pred_cam_t. And I printed out the min and max values of [x, y, z] of one person in demo video, which are: min[x, y, z] = [-2.5283, 0.3676, 134.07 ], max[x,y,z]=[2.5229, 0.9773, 243.8021]. So is it correct that z values times larger than x, y? Or can I just set the focal_length in base.py to a smaller one without retraining?