zhangmaoxiansheng / depthmap

Compiling

Environment: Linux opencv 4.0.0 cuda 10 Eigen3

B:two for optical flow (run_OF_*) and two for depth from stereo (run_DE_*). For each problem, a fast variant operating on intensity images (run_*_INT) and a slower variant operating on RGB images (run_*_RGB) is provided.

A:only run_DE_RGB

cd A(B)
mkdir build
cd build
cmake ../
make -j

Usage

The interface for all four binaries (run_*_*) is the same.

For depth map, run_DE_INT is faster

test video : test2.avi test2b.avi

./run_DE_INT test2.avi test2b.avi

output: A:background image,background depth B:mask,depth_mask

Parameters:(can be changed in depthmap.cpp)

1. Coarsest scale                               (here: 5)
2. Finest scale                                 (here: 3)
3/4. Min./Max. iterations                       (here: 12)
5./6./7. Early stopping parameters
8. Patch size                                   (here: 8)
9. Patch overlap                                (here: 0.4)
10.Use forward-backward consistency             (here: 0/no)
11.Mean-normalize patches                       (here: 1/yes)
12.Cost function                                (here: 0/L2)  Alternatives: 1/L1, 2/Huber, 10/NCC
13.Use TV refinement                            (here: 1/yes)
14./15./16. TV parameters alpha,gamma,delta     (here 10,10,5)
17. Number of TV outer iterations               (here: 1)
18. Number of TV solver iterations              (here: 3)
19. TV SOR value                                (here: 1.6)
20. Verbosity                                   (here: 2) Alternatives: 0/no output, 1/only flow runtime, 2/total runtime

The optical flow output is saves as .flo file. (http://sintel.is.tue.mpg.de/downloads)

The interface for depth from stereo is exactly the same. The output is saves as pfm file. (http://vision.middlebury.edu/stereo/code/)

NOTES:

1. For better quality, increase the number iterations (param 3/4), use finer scales (param. 2), higher patch overlap (param. 9), more outer TV iterations (param. 17)
2. L1/Huber cost functions (param. 12) provide better results, but require more iterations (param. 3/4)

The optical flow algorithm we use Fast DIS @inproceedings{kroegerECCV2016, Author = {Till Kroeger and Radu Timofte and Dengxin Dai and Luc Van Gool}, Title = {Fast Optical Flow using Dense Inverse Search}, Booktitle = {Proceedings of the European Conference on Computer Vision ({ECCV})}, Year = {2016}}