optflow: RLOF fixes (python-binding, status flag, flat textures) (#2097)
* * Bugfix for python-binding related to issue #2094. Copying the prevPtsMat to prevPoints was broken, when using python-binding.
* * Connect the blurred image correctly to CROSS support region segmentation. This allows to compute more robust support region shapes. Commit refers to issue #2094.
* * Bugfix to avoid the unintended assertion on flat textures discussed in issue #2094.
* * Bugfix on SparseRLOFOpticalFlow status was wrongly to to 0 if forward backward error < threshold, which is wrong. Inequation has been corrected. Refers to in issue #2087.
* small documentation changes
* bugfix on assertions and small test changes
Add robust local optical flow (RLOF) implementations (#1940)
* Add robust local optical flow (RLOF) implementations which is an improved pyramidal iterative Lucas-Kanade approach. This implementations contains interfaces for sparse optical flow for feature tracking and dense optical flow based on sparse-to-dense interpolation schemes.
Add performance and accuracy tests have been implementation as well as documentation with the related publications
* - exchange tabs with spaces
- fix optflow.bib indentation
- remove optflow_o.hpp
- change RLOFOpticalFlowParameter interfaces to Ptr<RLOFOpticalFlowParameter>
to remove error on building. Fix warnings
* introducing precompiler flag RLOD_SSE
* remove header that could not be found
* remove whitespaces
fix perf and accuracy tests
* remove x86intrin.h header
* fix ios and arm by removing last sse commands
* fix warnings for windows compilation
* fix documentation RLOFOpticalFlowParameter
* integrate cast to remove last warnings
* * add create method and function inferfaces to RLOFOpticalFlowParamter to enable python wrapper interfaces
* white space fixes / coding style
* fix perf test
* other changes: precomp.hpp / static
* use Matx44f and Vec4f instead of Mat
* normSigmas into constants
* replace ceil() calls
* maximum level is set to 5 so that it is similar value used in the papers
* implement paralellized horizontal cross segmentation as used in Geistert2016
* drop dead code
* Avoid using "data" and "step" calculations. Use .ptr<mat_type>(row, col) instead.
* Avoid using "data" and "step" calculations. Use .ptr<mat_type>(row, col) instead.
* bugfix on BEPLK with ica and adapt the accuracy tests
* more 'static' functions
* bugfix after changing ptr + step to .ptr(y,x) calls by adjusting ROI of
prevImage, currImage and derivI as well as changing the offset of the
points in the invoker classes.
* add some static_cast to avoid warning
* remove 50 grid size sample from perf test. This grid size is to sparse
for the epic interpolation
* remove notSameColor function since it is not used anymore
This patch adds ocl kernels to accelerate Dense Inverse Search
based optical flow algorithm, it acclerates 3 parts in the algorithm,
including 1) Structure tensor elements compute, 2) Patch inverse search,
3) Densification compute.
Perf and accuracy test are also added. The perf test shows it is 30%
faster than the current implementation.
Signed-off-by: Li Peng <peng.li@intel.com>
* Minor fixes
* Start adding correspondence finding
* Added finding of correspondences using GPC
* New evaluation tool for GPC
* Changed default parameters
* Display ground truth in the evaluation tool
* Added training tool for MPI Sintel dataset
* Added the training tool for Middlebury dataset
* Added some OpenCL optimization
* Added explanatory notes
* Minor improvements: time measurements + little ocl optimization
* Added demos
* Fixed warnings
* Make parameter struct assignable
* Fix warning
* Proper command line argument usage
* Prettified training tool, added parameters
* Fixed VS warning
* Fixed VS warning
* Using of compressed forest.yml.gz files by default to save space
* Added OpenCL flag to the evaluation tool
* Updated documentation
* Major speed and memory improvements:
1) Added new (optional) type of patch descriptors which are much faster. Retraining with option --descriptor-type=1 is required.
2) Got rid of hash table for descriptors, less memory usage.
* Fixed various floating point errors related to precision.
SIMD for dot product, forest traversing is a little bit faster now.
* Tolerant floating point comparison
* Triplets
* Added comment
* Choosing negative sample among nearest neighbors
* Fix warning
* Usage of parallel_for_() in critical places. Performance improvments.
* Simulated annealing heuristic
* Moved OpenCL kernel to separate file
* Moved implementation to source file
* Added basic accuracy tests for GPC and PCAFlow
* Fixing warnings
* Test accuracy constraints were too strict
* Test accuracy constraints were too strict
* Make tests more lightweight
Added variational refinement as a separate class (based on implementation
inside DeepFlow, but significantly accelerated, about 4-6 times faster),
accelerated the main dense inverse search algorithm. Added several new
features including patch mean normalization for increased robustness to
illumination changes and spatial propagation, which often helps to recover
from errors introduced by the coarse-to-fine scheme. Expanded the
documentation, added new accuracy and perf tests. Refactored some of
the already existing optical flow accuracy tests.
* Sparse match interpolator interface and EdgeAwareInterpolator were
added to the ximgproc module
* New optical flow algorithm, based on PyrLK sparse OF and sparse match
interpolation, is added to the optflow module
Tobias Senst
Vadim Pisarevsky
Alexander Alekhin
Yuriy Obukh
Li Peng
Vladislav Samsonov
Alexander Bokov