SfMLearner
SfMLearner copied to clipboard
tinghuiz
An unsupervised learning framework for depth and ego-motion estimation from monocular videos
Like the following picture.The network seems to learn the structure well,but the distance is not learned well.The depth in the distance seems to be similar to that in the near...
HI, Thank you for your work. When I add ssim loss to the loss function, the depth map gets a strange black border result. The result is as follows. 
Here is your readme‘s results | Abs Rel | Sq Rel | RMSE | RMSE(log) | Acc.1 | Acc.2 | Acc.3 | |---------|--------|-------|-----------|-------|-------|-------| | 0.183 | 1.595 | 6.709 |...
In this paper,  there are 44540 sequences in total after running "prepare_train_data.py". But I run this code, I got 45028 sequences.The number of training sequences is 40250 and validate...
I'm getting a FileNotFoundError when I run `test_kitti_pose.py`. It seems that the dataset downloaded by `download_kitti_pose_eval_data.sh` doesn't have the expected structure: `test_kitti_pose.py` is looking for a file **${POSE_DATASET}/sequences/09/times.txt**. But the...
I have downloaded the KITTI odometry gray scale dataset. When I'm trying to run the pose model on the dataset, I'm getting the "InvalidArgumentError (see above for traceback): Assign requires...
Hi,tinghui. How to split the groundtruth pose into interframe pose? Can you give the code to me please?
When testing the pose prediction accuracy (in `pose_evaluation_utils.py`), the ATE of a snippet is divided by the snippet length. Both predicted and GT snippets have the Identity transformation as their...
how can you connect 5-frame snippets (with the format of tum, (timestamp tx ty tz qx qy qz qw) into a complete pose estimate in a global coordinate system?
I want to see the accuracy of the warping. So how can i see the warped image during the training stage. I want to make sure whether the source image...
