On Building an Accurate Stereo Matching System on Graphics Hardware

Slides:



Advertisements
Similar presentations
Discontinuity Preserving Stereo with Small Baseline Multi-Flash Illumination Rogerio Feris 1, Ramesh Raskar 2, Longbin Chen 1, Karhan Tan 3 and Matthew.
Advertisements

Efficient High-Resolution Stereo Matching using Local Plane Sweeps Sudipta N. Sinha, Daniel Scharstein, Richard CVPR 2014 Yongho Shin.
875: Recent Advances in Geometric Computer Vision & Recognition
Spatial-Temporal Consistency in Video Disparity Estimation ICASSP 2011 Ramsin Khoshabeh, Stanley H. Chan, Truong Q. Nguyen.
Stereo Vision Reading: Chapter 11
Stereo Many slides adapted from Steve Seitz. Binocular stereo Given a calibrated binocular stereo pair, fuse it to produce a depth image Where does the.
Cuong Cao Pham and Jae Wook Jeon, Member, IEEE
M.S. Student, Hee-Jong Hong
Real-Time Accurate Stereo Matching using Modified Two-Pass Aggregation and Winner- Take-All Guided Dynamic Programming Xuefeng Chang, Zhong Zhou, Yingjie.
Stereo Matching Segment-based Belief Propagation Iolanthe II racing in Waitemata Harbour.
Does Color Really Help in Dense Stereo Matching?
2009/04/07 Yun-Yang Ma.  Overview  What is CUDA ◦ Architecture ◦ Programming Model ◦ Memory Model  H.264 Motion Estimation on CUDA ◦ Method ◦ Experimental.
Boundary matting for view synthesis Samuel W. Hasinoff Sing Bing Kang Richard Szeliski Computer Vision and Image Understanding 103 (2006) 22–32.
Last Time Pinhole camera model, projection
Contents Description of the big picture Theoretical background on this work The Algorithm Examples.
1 Static Sprite Generation Prof ︰ David, Lin Student ︰ Jang-Ta, Jiang
Stereo & Iterative Graph-Cuts Alex Rav-Acha Vision Course Hebrew University.
High-Quality Video View Interpolation
The plan for today Camera matrix
Virtual Control of Optical Axis of the 3DTV Camera for Reducing Visual Fatigue in Stereoscopic 3DTV Presenter: Yi Shi & Saul Rodriguez March 26, 2008.
Stereo Computation using Iterative Graph-Cuts
Binocular Stereo Philippos Mordohai University of North Carolina at Chapel Hill September 21, 2006.
Aleixo Cambeiro Barreiro 광주과학기술원 컴퓨터 비전 연구실
CSE473/573 – Stereo Correspondence
A Performance and Energy Comparison of FPGAs, GPUs, and Multicores for Sliding-Window Applications From J. Fowers, G. Brown, P. Cooke, and G. Stitt, University.
IEEE TRANSACTIONS ON IMAGE PROCESSING, VOL. 20, NO. 11, NOVEMBER 2011 Qian Zhang, King Ngi Ngan Department of Electronic Engineering, the Chinese university.
Announcements PS3 Due Thursday PS4 Available today, due 4/17. Quiz 2 4/24.
Proceedings of the 2004 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR’04) /04 $20.00 c 2004 IEEE 1 Li Hong.
Joint Histogram Based Cost Aggregation For Stereo Matching Dongbo Min, Member, IEEE, Jiangbo Lu, Member, IEEE, Minh N. Do, Senior Member, IEEE IEEE TRANSACTION.
Introduction Belief propagation: known to produce accurate results for stereo processing/ motion estimation High storage requirements limit the use of.
Stereo Matching Information Permeability For Stereo Matching – Cevahir Cigla and A.Aydın Alatan – Signal Processing: Image Communication, 2013 Radiometric.
Michael Bleyer LVA Stereo Vision
A Rapid Stereo Matching Algorithm Based on Disparity Interpolation Gang Yao Yong Liu Bangjun Lei Dong Ren Institute of Intelligent Vision and Image Information.
Mutual Information-based Stereo Matching Combined with SIFT Descriptor in Log-chromaticity Color Space Yong Seok Heo, Kyoung Mu Lee, and Sang Uk Lee.
A Local Adaptive Approach for Dense Stereo Matching in Architectural Scene Reconstruction C. Stentoumis 1, L. Grammatikopoulos 2, I. Kalisperakis 2, E.
Joint Depth Map and Color Consistency Estimation for Stereo Images with Different Illuminations and Cameras Yong Seok Heo, Kyoung Mu Lee and Sang Uk Lee.
Rohith MV, Gowri Somanath, Chandra Kambhamettu Video/Image Modeling and Synthesis(VIMS) Lab, Dept. of Computer and Information Sciences Cathleen Geiger.
Takuya Matsuo, Norishige Fukushima and Yutaka Ishibashi
Object Stereo- Joint Stereo Matching and Object Segmentation Computer Vision and Pattern Recognition (CVPR), 2011 IEEE Conference on Michael Bleyer Vienna.
Cross-Based Local Multipoint Filtering
A Non-local Cost Aggregation Method for Stereo Matching
On a Few Ray Tracing like Algorithms and Structures. -Ravi Prakash Kammaje -Swansea University.
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE, VOL. 34, NO. 2, FEBRUARY Leonardo De-Maeztu, Arantxa Villanueva, Member, IEEE, and.
Stereo Many slides adapted from Steve Seitz.
Feature-Based Stereo Matching Using Graph Cuts Gorkem Saygili, Laurens van der Maaten, Emile A. Hendriks ASCI Conference 2011.
Computer Vision, Robert Pless
A Region Based Stereo Matching Algorithm Using Cooperative Optimization Zeng-Fu Wang, Zhi-Gang Zheng University of Science and Technology of China Computer.
1 Real-Time Stereo-Matching for Micro Air Vehicles Pascal Dufour Master Thesis Presentation.
Window-based Approach For Fast Stereo Correspondence Raj Kumar Gupta, Siu-Yeung Cho IET Computer Vision,
Computer Vision Lecture #10 Hossam Abdelmunim 1 & Aly A. Farag 2 1 Computer & Systems Engineering Department, Ain Shams University, Cairo, Egypt 2 Electerical.
Fast Census Transform-based Stereo Algorithm using SSE2
Segmentation of Vehicles in Traffic Video Tun-Yu Chiang Wilson Lau.
Solving for Stereo Correspondence Many slides drawn from Lana Lazebnik, UIUC.
Improved Census Transforms for Resource-Optimized Stereo Vision
Jeong Kanghun CRV (Computer & Robot Vision) Lab..
Visual Odometry David Nister, CVPR 2004
Journal of Visual Communication and Image Representation
Efficient Stereo Matching Based on a New Confidence Metric
Project 2 due today Project 3 out today Announcements TexPoint fonts used in EMF. Read the TexPoint manual before you delete this box.: AAAAA.
Canny Edge Detection Using an NVIDIA GPU and CUDA Alex Wade CAP6938 Final Project.
Electrical Engineering Department of National Cheng Kung University Smart Media & Intelligent Living Excellence Lab /08/14 指導教授:詹寶珠教授 報告者:陳正 旻 期末.
Stereo Video 1. Temporally Consistent Disparity Maps from Uncalibrated Stereo Videos 2. Real-time Spatiotemporal Stereo Matching Using the Dual-Cross-Bilateral.
數位三維視訊 楊 家 輝 Jar-Ferr Yang 電腦與通信工程研究所 電機工程學系 國立成功大學 Institute of Computer and Communication Engineering Department of Electrical Engineering National Cheng.
1 Munther Abualkibash University of Bridgeport, CT.
Local Stereo Matching Using Motion Cue and Modified Census in Video Disparity Estimation Zucheul Lee, Ramsin Khoshabeh, Jason Juang and Truong Q. Nguyen.
Summary of “Efficient Deep Learning for Stereo Matching”
Semi-Global Matching with self-adjusting penalties
Jure Zbontar, Yann LeCun
SoC and FPGA Oriented High-quality Stereo Vision System
Computing the Stereo Matching Cost with a Convolutional Neural Network
Presentation transcript:

On Building an Accurate Stereo Matching System on Graphics Hardware Xing Mei ; Xun Sun ;  Mingcai Zhou ;  Shaohui Jiao ;   Haitao Wang ; Xiaopeng Zhang  Samsung Advanced Institute of Technology, China Lab Computer Vision Workshops, 2011 IEEE

Outline Introduction Related Works Algorithmn CUDA Implementation Experimental Results Conclusion

Introduction

Introduction Dense two-frame stereo matching Compute a disparity map from stereo images. Broad applications: 3D reconstruction, view interpolation

Related Works

Related Works Local methods Global methods Compute each pixel’s disparity independently over a local support region. Fast but inaccurate. Global methods Solve the stereo problem in an energy minimization process. Accurate but slow due to time-comsuming global optimizer.(GC,BP)

Related Works Propagation-based methods Produce quasi-dense or dense disparity results from a set of seed pixels. Relatively fast but sensitive to early wrong matches use segmented regions as guided propagation unit expensive cost

Related Works Introduce a simple guided unit for propagation : pixel-wise 1D line segments. No image segmentation required here. Simple, fast and accurate

Algorithmn

Algorithmn AD-Census Cost Initialization Framework Multi-step Disparity Refinement Scanline Optimization Cross-based Cost Aggregation AD-Census Cost Initialization Input: Stereo images Output: Disparity map

Algorithmn AD-Census Cost Initialization Cross-based Cost Aggregation Multi-step Disparity Refinement Scanline Optimization Cross-based Cost Aggregation AD-Census Cost Initialization Input: Stereo images Output: Disparity map

Disparity Cost Computing Cost mesure : AD, BT, gradient-based measures, non-parametric transforms(rank/census[3])...... Combination : SAD + gradient[6] , AD + Census AD (Absolute Distance) Constant color assumption Repetitive structures Census Encodes local image structures Textureless regions [3] H. Hirschmuller and D. Scharstein. “Evaluation of stereo matching costs on images with radiometric differences.”IEEE TPAMI, 31(9):2009. [6] A. Klaus, M. Sormann, and K. Karner. “Segment-based stereo matching using belief propagation and a self-adapting dissimilarity measure.” ICPR,2006.

AD-Census Cost Initialization 𝐶 p,𝑑 = 𝜌( 𝐶 𝑐𝑒𝑛𝑠𝑢𝑠 p,𝑑 , 𝜆 𝑐𝑒𝑛𝑐𝑢𝑠 )+𝜌( 𝐶 𝐴𝐷 p,𝑑 , 𝜆 𝐴𝐷 ) p : pixel d : level >> a robust function on variable 𝑐 pd = (x-d,y) in the right image 𝐶 𝑐𝑒𝑛𝑠𝑢𝑠 : Hamming distance[22] d Left I Right I [22] R. Zabih and J. Woodfill. “Non-parametric local transforms for computing visual correspondence.” In Proc. ECCV, 1994.

Census Transform Census transform window : 121 130 26 31 39 109 115 33 40 30 98 102 78 67 45 47 32 170 198 86 99 159 210 1 X Census transform window : 1

Census Hamming Distance Left image Right image Hamming Distance = 3 1 XOR 1 1

AD-Census Cost Initialization 𝐶 p,𝑑 = 𝜌( 𝐶 𝑐𝑒𝑛𝑠𝑢𝑠 p,𝑑 , 𝜆 𝑐𝑒𝑛𝑐𝑢𝑠 )+𝜌( 𝐶 𝐴𝐷 p,𝑑 , 𝜆 𝐴𝐷 ) > >> a robust function on variable 𝑐

AD-Census Cost Initialization AD-Census measure produces proper disparity results for both repetitive structures and textureless regions.

Algorithmn AD-Census Cost Initialization Cross-based Cost Aggregation Multi-step Disparity Refinement Scanline Optimization Cross-based Cost Aggregation AD-Census Cost Initialization Input: Stereo images Output: Disparity map

Cross-based Cost Aggregation[23] Cross construction Line ending points P1, P2 for P are located when rule 1 or 2 are violated: R1: Color self-similarity in the line region: smooth depth assumption R2: Arm length limitation: avoid over-smoothness [23] K. Zhang, J. Lu, and G. Lafruit. “Cross-based local stereo matching using orthogonal integral images.” IEEE TCSVT,2009.

Cross-based Cost Aggregation

Cross-based Cost Aggregation Enhance cross construction (use pixel p’s left arm and the endpoint pixel pl as an example)

Cross-based Cost Aggregation Run this step for 4 iterations to get stable cost values. For iteration 1 and 3, aggregated horizontally and then vertically. For iteration 2 and 4, aggregated vertically and then horizontally. Reduce the errors at depth discontinuities.

Cross-based Cost Aggregation Our aggregation method can better handle large textureless regions and depth discontinuities.

Cross-based Cost Aggregation [21] K.-J. Yoon and I.-S. Kweon. “Adaptive support-weight approach for correspondence search.” IEEE TPAMI, 2006. [23] K. Zhang, J. Lu, and G. Lafruit. “Cross-based local stereo matching using orthogonal integral images.” IEEE TCSVT,2009.

Algorithmn AD-Census Cost Initialization Cross-based Cost Aggregation Multi-step Disparity Refinement Scanline Optimization Cross-based Cost Aggregation AD-Census Cost Initialization Input: Stereo images Output: Disparity map

Scanline Optimization[2] 4 scanline optimization processes are performed independently. 2 horizontal directions 2 vertical directions 𝐶 2 𝐶 𝑟 [2] H. Hirschmuller. Stereo processing by semiglobal matching and mutual information.” IEEE TPAMI, 2008.

Scanline Optimization p-r p r r : direction p-r : the previous pixel along the same direction 𝑃1, 𝑃2 : penalize the disparity changes between neighboring pixels. (𝑃1 ≤ 𝑃2 ) [8] [8]S. Mattoccia, F. Tombari, and L. D. Stefano. “Stereo vision enabling precise border localization within a scanline optimization framework.” In Proc. ACCV, pages 517–527, 2007.

Scanline Optimization The final cost : The disparity with the minimum 𝐶2 value is selected as pixel p’s intermediate result. 𝐶 2 𝐶 𝑟

Algorithmn AD-Census Cost Initialization Cross-based Cost Aggregation Multi-step Disparity Refinement Scanline Optimization Cross-based Cost Aggregation AD-Census Cost Initialization Input: Stereo images Output: Disparity map

Multi-step Disparity Refinement Outlier Handling Outlier Detection Iterative Region Voting Proper Interpolation Depth Discontinuity Adjustment Sub-pixel Enhancement

Outlier Handling--Detection The outliers : 𝐷𝐿(p) != 𝐷R(p − (𝐷𝐿(p), 0)) Outliers are further classified into occlusion and mismatch points p intersect its epipolar line and 𝐷R is checked If no intersection p is labelled as “occlusion”, otherwise “mismatch”

Outlier Handling--Iterative Region Voting Construct cross-based regions and a robust voting scheme Sp : 𝜏𝑆, 𝜏𝐻 : threshold values 5 iterations d d

Outlier Handling--Proper Interpolation occlusion The pixel with the lowest disparity value is selected for interpolation It’s most likely comes from the background mismatch points The pixel with the most similar color is selected for interpolation.

Depth Discontinuity Adjustment For each pixel p on the disparity edge, two pixels p1, p2 from both sides of the edge are collected. 𝐷𝐿(p) is replaced by 𝐷𝐿(p1) or 𝐷𝐿(p2) if one of the two pixels has smaller matching cost than 𝐶2(p,𝐷𝐿(p)). 𝐷𝐿(P1) 𝐷𝐿(P) 𝐷𝐿(P2)

Sub-pixel Enhancement[20] Quadratic polynomial interpolation With 3*3 median filter [20] Q. Yang, L. Wang, R. Yang, H. Stewenius, and D. Nister. “Stereo matching with color-weighted correlation, hierarchical belief propagation and occlusion handling.” IEEE TPAMI, 2009.

Multi-step Disparity Refinement The average error percentages after performing each refinement step.

CUDA Implementation

CUDA Implementation Compute Unified Device Architecture (CUDA) is a programming interface for parallel computation tasks on NVIDIA graphics hardware. The computation task is coded into a kernel function. The allocation of the threads is controlled with two hierarchical concepts: grid and block. A kernel creates a grid with multiple blocks, and each block consists of multiple threads. Kernel Grid Block Thread …

CUDA Implementation Cost Initialization: Parallelize with 𝑊 × 𝐻 threads. Organize into a 2D grid and the block size is set to 32× 32. Each thread computes a cost value for a pixel at a given disparity. For census transform, a square window is require for each pixel, which requires loading more data into the shared memory for fast access. Kernel Grid Block Thread … 32X32

CUDA Implementation Cross-based Cost Aggregation: A grid with 𝑊 × 𝐻 threads. Cross construction : block size is 𝑊 or 𝐻 to efficiently handle a scan line Cost aggregation : block size is 32X32 Data reuse with shared memory is considered in both steps.

CUDA Implementation Scanline Optimization: Disparity Refinement: This step is different, because the process is sequential in the scanline direction and parallel in the orthogonal direction. 𝑊 × 𝐷 or 𝐻 × 𝐷 threads Disparity Refinement: 𝑊 × 𝐻 threads

Experimental Results

Experimental Results Device : A PC with Core 2 Duo 2.20GHz CPU and NVIDIA GeForce GTX 480 graphics card Settings parameters: Source : Middlebury http://vision.middlebury.edu/stereo/ HHI database(book arrival) Microsofy i2i database(Ilkay)

Experimental Results Tsukuba Venus Teddy Cones CPU 2.5 4.5 15 GPU 0.015 0.032 0.095 0.094 The GPU-friendly system brings an impressive 140× speedup. The average proportions of the GPU running time for the four computation steps are 1%, 70%, 28% and 1% respectively. The iterative cost aggregation step and the scanline optimization process dominate the running time.

Experimental Results First row: disparity maps generated with our system. Second row: disparity error maps with threshold 1. Errors in unoccluded and occluded regions are marked in black and gray respectively.

Experimental Results

Experimental Results video

Snapshots on ’book arrival’ stereo video Experimental Results Snapshots on ’book arrival’ stereo video

Experimental Results Snapshots on ’Ilkay’ stereo video

Conclusion

Conclusion Contributions Future works Present a near real-time stereo system with accurate disparity results. Combine some known techniques without sacrificing performance and parallelism to obtain the high quality disparity map. Future works Improve to apply it in real world applications Robust parameter setting methods