StereoSGBM algorithm updated to use wide universal intrinsics#15478
Merged
opencv-pushbot merged 1 commit intoopencv:3.4from Oct 31, 2019
Merged
StereoSGBM algorithm updated to use wide universal intrinsics#15478opencv-pushbot merged 1 commit intoopencv:3.4from
opencv-pushbot merged 1 commit intoopencv:3.4from
Conversation
Contributor
Author
Performance for SSE2 baseline
Performance for SSE3 baseline
Performance for SSE4_2 baseline
Performance for AVX2 baseline
|
3be502c to
d9d62fb
Compare
alalek
reviewed
Oct 22, 2019
Member
alalek
left a comment
There was a problem hiding this comment.
There is crashed test on AVX512: Calib3d_StereoSGBM.regression
Please take a look.
| C[x] = (CostType)(Cprev[x] + hsumAdd[x] - hsumSub[x]); | ||
| #endif | ||
| } | ||
| else*/ |
Contributor
Author
There was a problem hiding this comment.
This part perform cost evaluation for the bottom of the image, but it was missed in the original implementation. So result of the code is inconsistent with test reference data. However bottom evaluation works inaccurate anyway due to matching of mostly border padding instead of real data.
modules/calib3d/src/stereosgbm.cpp
Outdated
| v_store_aligned(cost + x*D + d, _c0 + v_reinterpret_as_s16(diff1 >> diff_scale)); | ||
| v_store_aligned(cost + x*D + d + 8, _c1 + v_reinterpret_as_s16(diff2 >> diff_scale)); | ||
| } | ||
| for( ; d <= maxD - v_uint8::nlanes; d += v_uint8::nlanes ) |
Member
There was a problem hiding this comment.
v_uint8::nlanes
Why is v_int16::nlanes * 2 not used? (as destination buffer access is performed by two int16 vectors)
| for( x = width-1-maxX2; x < width-1- minX2; x++ ) | ||
| // to process values from [minX2, maxX2) we should check memory location (width - 1 - maxX2, width - 1 - minX2] | ||
| // so iterate through [width - maxX2, width - minX2) | ||
| for( x = width-maxX2; x < width-minX2; x++ ) |
Member
There was a problem hiding this comment.
prow2[x] indexes are changed - intentionally?
d9d62fb to
71ffd6a
Compare
71ffd6a to
42b1d04
Compare
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
resolves #15206
This pullrequest changes
StereoSGBM algorithm updated to use wide universal intrinsics