Wenjing Bian

Results 29 comments of Wenjing Bian

Hi, please let me know if this issue still exists

As the model is trained using ShapeNet with white background, you will need to mask the image during inference.

Can you suggest what data you used? Training on more shapenet categories may help if your custom data is of similar categories.

Hi, if you still have this problem, please give me your email address, I can share the preprocessed data with you through email.

> @bianwenjing Hi, I am having problems preprocessing the data as well, could you please share the preprocessed data via email to [[email protected]](mailto:[email protected])? Can you check if you received my...

Hi, we do not support batch_size >1 at this moment. Sorry for the inconvenience.

Hi, the intrinsics loaded into the model is a 3*3 numpy array with the form [[fx, 0, cx], [0, fy, cy], [0, 0, 1]]. You can customise it [here](https://github.com/ActiveVisionLab/nope-nerf/blob/1b3905776b461539e1d71493a3ac916d2f00aeaa/dataloading/dataset.py#L84)

你好,你可以使用[LLFF](https://github.com/Fyusion/LLFF)的代码通过colmap直接生成poses_bounds.npy文件。

你好,生成的poses_bounds.npy中已经包含intrinsic,不需要另外获取intrinsics.npz

你好,你可以使用`vis/render.py` 生成图片,深度图及其视频,请参考`README.md`. 我们目前没有提供点云的渲染