High accuracy local stereo matching using DoG scale map

Kitagawa, Masamichi; Shimizu, Ikuko; Šára, Radim

doi:10.23919/mva.2017.7986850

Cited by 27 publications

(11 citation statements)

References 3 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The results in Table 7 show that our approach gives average absolute errors in non-occluded regions equal to 9.76% and average absolute errors in all regions equal to 17.6%. The total execution time of our proposed algorithm is 0.27 s. These results indicate that our approach is more accurate and faster than DoGGuided [33], BSM [34], ICSG [35], and SGBM1 [36]. SPS [32] produces more accurate results over all regions (16.6%) but its results are less accurate in non-occluded regions (10.4%).…”

Section: Implementation Resultsmentioning

confidence: 84%

“…The obtained results are detailed in Tables 4 and 5, which respectively show the error for each stereo pair on non occluded and all regions. The evaluation on MV3 indicates that our proposed approach outperforms other algorithms such as DoGGuided [33] that use a guided filter based on the response of the difference of Gaussian, binary stereo matching (BSM) [34] and other recent approaches. In addition our stereo matching algorithm is more accurate than ICSG [35] and semi global matching (SGBM1) [36].…”

Section: Experimental Results and Analysismentioning

confidence: 98%

See 1 more Smart Citation

Real-time stereo matching on CUDA using Fourier descriptors and dynamic programming

Hammami

Smach²,

Atri

2019

Comp. Visual Media

View full text Add to dashboard Cite

Computation of stereoscopic depth and disparity map extraction are dynamic research topics. A large variety of algorithms has been developed, among which we cite feature matching, moment extraction, and image representation using descriptors to determine a disparity map. This paper proposes a new method for stereo matching based on Fourier descriptors. The robustness of these descriptors under photometric and geometric transformations provides a better representation of a template or a local region in the image. In our work, we specifically use generalized Fourier descriptors to compute a robust cost function. Then, a box filter is applied for cost aggregation to enforce a smoothness constraint between neighboring pixels. Optimization and disparity calculation are done using dynamic programming, with a cost based on similarity between generalized Fourier descriptors using Euclidean distance. This local cost function is used to optimize correspondences. Our stereo matching algorithm is evaluated using the Middlebury stereo benchmark; our approach has been implemented on parallel high-performance graphics hardware using CUDA to accelerate our algorithm, giving a real-time implementation.

show abstract

Section: Implementation Resultsmentioning

confidence: 84%

Section: Experimental Results and Analysismentioning

confidence: 98%

Real-time stereo matching on CUDA using Fourier descriptors and dynamic programming

Hammami

Smach²,

Atri

2019

Comp. Visual Media

View full text Add to dashboard Cite

show abstract

“…In addition, it can be seen that our algorithm also achieves better performance than ISM [19], and our average error rate as well as the average disparity error are respectively 10.8% and 2.81 px lower than ISM. Obviously, our method also outperforms the remaining four non-learning stereo methods such as DoGGuided [45] by a large margin. As illustrated in Fig.4, Bicycle2, Classroom2, Computer, Crusade and Newkuba in the test set of the Middlebury dataset v3 are used for visual comparison.…”

Section: B Evaluation On Middlebury Dataset V3mentioning

confidence: 86%

Stereo Matching With Fusing Adaptive Support Weights

Zhu

et al. 2019

IEEE Access

View full text Add to dashboard Cite

Stereo matching has been widely used in various computer vision applications and it is still a challenging problem. Adaptive support weights (ASW) methods represent the state of the art in stereo matching and have achieved outstanding performance. However, the local ASW methods fail to resolve the matching ambiguity in low texture areas because their cost aggregation is limited within local fixed or adaptive support windows. On the other hand, the non-local ASW methods perform cost aggregation along a special tree, so that these methods are often sensitive to high texture areas since some useful connectivity constrains between adjacent pixels are broken during constructing the special tree. To solve these problems, in this paper, a novel and generic fusing ASW framework are proposed for stereo matching. In this framework, we establish dual support windows for each pixel, i.e., a local window and the whole image window. As such, the primitive connectivity between each pixel and its neighboring pixels in the local window can be maintained, and then each pixel not only gets appropriate supports from neighboring pixels within its local support window but also receives more adaptive supports from the other pixels outside the local window. Furthermore, a local edge-aware filter and a non-local edge-aware filter, whose kernel windows correspond to the dual support windows, are merged in order to achieve collaborative filtering of the cost volume. The performance evaluation on the Middlebury and KITTI datasets shows that the proposed stereo matching method outperforms the current state-of-the-art methods. INDEX TERMS Stereo matching, cost aggregation, adaptive support weight, edge-aware filtering.

show abstract

“…We have compared our results with recently published methods (i.e. [4–8 ]) to show the competitiveness of the proposed method in this Letter. Their method were developed with different framework architectures including the deep learning method in [9 ].…”

Section: Introductionmentioning

confidence: 94%

Improvement of disparity map refinement stage using adaptive least square plane fitting technique

Hamzah

Ibrahim

2020

Electronics Letters

View full text Add to dashboard Cite

This Letter presents an improvement of disparity map refinement stage using adaptive least square plane fitting technique. This technique is proposed to increase the accuracy on the final stage of stereo matching algorithm. Fundamentally, the accuracy of matching process depends on the robustness of an algorithm on the plain colour region, depth discontinuity and repetitive pattern area. These regions are difficult to be matched and very challenging. Thus, this Letter proposes multiple point selections of disparities on a plane to adaptively group similar disparity values. Then, the distance of each disparity is calculated to ensure the values are accurately grouped. The distance refers to an assigned reference radius disparity location. Based on the experimental results using the KITTI and Middlebury benchmark images, the proposed method is capable of increasing the accuracy on the challenging regions and the quantitative measurement.

show abstract

High accuracy local stereo matching using DoG scale map

Cited by 27 publications

References 3 publications

Real-time stereo matching on CUDA using Fourier descriptors and dynamic programming

Real-time stereo matching on CUDA using Fourier descriptors and dynamic programming

Stereo Matching With Fusing Adaptive Support Weights

Improvement of disparity map refinement stage using adaptive least square plane fitting technique

Contact Info

Product

Resources

About