Jiawang Bian
Jiawang Bian
The reported results are by DispResNet...
> same problem！ Do you try the latest version?
Odometry is an accumulated sequentially from pairwise predictions, so the results using different models would be different even though these models show very similar accuracy on pairwise prediction.
It is similar to train on the NYUv2 dataset. 1. You need to downsample the video. For example, pick one image from every 10 images. 2. You may train the...
try infinitamv3
In terms of implementation, it is OK. However, I am not sure whether the single view depth network can be learned using fisheye camera. Looking forward to your results.
1. The first problem may be little device difference, as you can see that the results are very similar. You can report your reproduced results in future paper for fair...
auto_mask is at Line 104 of "loss_functions.py", and Ms is at Line 112 of "loss_functions.py". You need to write a script to save the results for visualization.
See "https://github.com/JiawangBian/Unsupervised-Indoor-Depth" for details. You can download the dataset there.
Thanks for your attention. The code is based on sfmleaner pytorch, where with-gt is used. However, we find this is important for sfmlearner but not so critical to our method,...