Andyen512

训练自己的数据集

5

请问如果我自己的数据集中并不是按照SMPL的24个关键点，比如我有26个关节，与SMPL中的24个不同，可以用您的方法生成最后的旋转参数poses吗

KeyError: 'trans'

3

when I ues my own npz file, it got wrongs as follows, can you tell me why? in run self._target(*self._args, **self._kwargs) File "server.py", line 26, in tcplink data = getData()...

Hi chungyi, I know the 'motion_weights_priors' consists of ellipsoidal 3D Gaussian around each body bone. but I'm not understand the implementation of `dist = np.einsum('abci, abci->abc', np.einsum('abci, ij->abcj', grid, SIGMA),...

SMPL-X model

1

Hi chungyi, I'd like to ask does humannerf support SMPL-X model?

Thanks for the great work! I'd like to know what's the difference between the Sample Field S and Se, is it just different in the number of FC layers? Using...

Thanks for your great work! If possible, could you please share the visualization code for drawing figure 8 in the paper? Thank you very much!

请问出现以下的错误的原因是什么呢 best_ious, best_n = ious.max(0) # get each gt box matched best anchor RuntimeError: cannot perform reduction function max on tensor with no elements because the operation does not have...

关于calib矩阵的一些问题

3

您好，最近刚入门这个方向，在看PIFu的时候发现了您的仓库，有一些相关的问题想请教一下： （1）在PIFu的[TrainDataset.py](https://github.com/shunsukesaito/PIFu/blob/f0a9c99ef887e1eb360e865a87aa5f166231980e/lib/data/TrainDataset.py#L139)中，从生成的PARAM中提取出了ortho_ratio、scale、center参数，并利用这些参数计算得到了scale_intrinsic、uv_intrinsic，并最终算出了calib这样一个变换矩阵，但我并不是很懂calib的含义以及这个变换的过程，请您指点一下，或是能否给一些参考链接我去学习一下。 （2）第二个问题与（1）也有些相关，是在[HGPIFuNet.py](https://github.com/shunsukesaito/PIFu/blob/f0a9c99ef887e1eb360e865a87aa5f166231980e/lib/model/HGPIFuNet.py#L85)中，为什么某一个三维坐标值通过calib矩阵做投影之后，得到的z坐标值就是对应该点的深度值？ 希望得到您的回复。

Hi, Thanks for your great work where is the depth initialization in the code

Video-based experiments code

1

Hi, When the code of Video-based experiments will be released?