Hardware-efficient belief propagation

Liang, Chia-Kai; Cheng, Chao-Chung; Lai, Yen-Chieh; Chen, Liang-Gee; Chen, Homer H.

doi:10.1109/cvpr.2009.5206819

Cited by 26 publications

(27 citation statements)

References 34 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…This is, however, hard to extend to large SIMD-like architectures, where deactivation is preferably done per block. Overlapping tiles have also been proposed [71,73], with most computations done locally within each tile. However, due to the local memory required for storing of messages, one becomes limited to relatively small tiles, which makes it harder to keep all SPs fully active.…”

Section: Mrf Inferencementioning

confidence: 99%

“…One further needs to propagate messages between tiles, which means that multiple kernel calls are still required, with propagated messages stored to global memory. Since a hierarchical framework reduces the number of iterations per level anyway, its hard to simultaneously exploit the joint benefits of both local tile processing and hierarchies [73].…”

Section: Mrf Inferencementioning

confidence: 99%

See 1 more Smart Citation

Detecting, segmenting and tracking unknown objects using multi-label MRF inference

Björkman

Bergström

Kragić

2014

Computer Vision and Image Understanding

View full text Add to dashboard Cite

This article presents a unified framework for detecting, segmenting and tracking unknown objects in everyday scenes, allowing for inspection of object hypotheses during interaction over time. A heterogeneous scene representation is proposed, with background regions modeled as a combinations of planar surfaces and uniform clutter, and foreground objects as 3D ellipsoids. Recent energy minimization methods based on loopy belief propagation, tree-reweighted message passing and graph cuts are studied for the purpose of multi-object segmentation and benchmarked in terms of segmentation quality, as well as computational speed and how easily methods can be adapted for parallel processing. One conclusion is that the choice of energy minimization method is less important than the way scenes are modeled. Proximities are more valuable for segmentation than similarity in colors, while the benefit of 3D information is limited. It is also shown through practical experiments that, with implementations on GPUs, multiobject segmentation and tracking using state-of-art MRF inference methods is feasible, despite the computational costs typically associated with such methods.

show abstract

Section: Mrf Inferencementioning

confidence: 99%

Section: Mrf Inferencementioning

confidence: 99%

Detecting, segmenting and tracking unknown objects using multi-label MRF inference

Björkman

Bergström

Kragić

2014

Computer Vision and Image Understanding

View full text Add to dashboard Cite

show abstract

“…Five of up-to-date top-10-ranked algorithms (with default error threshold equals to 1) are based on the optimized global energy function [15]. Although global methods can reach a high quality level with VGA@30 frames per second (fps) performance [16][17], it is still hard for real-time and high resolution application cases because of its computation complexity. Local approach is based on color or intensity patterns within a finite window to determine the disparity.…”

Section: Stereo Matching Algorithms Overviewmentioning

confidence: 99%

Software and Hardware Implementations of Stereo Matching

Zhou¹,

Sun²,

Zhan³

et al. 2014

IJSIP

View full text Add to dashboard Cite

show abstract

“…In order to achieve real-time, recent advances exploit the parallel computational power within GPUs [12,13]. With some necessary modifications, some global methods also implemented in the very large scale integration circuit [14,15] at the expense of considerable high consumption of logic source, memory and bandwidth.…”

Section: Background and Related Workmentioning

confidence: 99%

Real-time stereo vision system using adaptive weight cost aggregation approach

Ding

Liu

Zhou

et al. 2011

J Image Video Proc.

View full text Add to dashboard Cite

Many vision applications require high-accuracy dense disparity maps in real time. Due to the complexity of the matching process, most real-time stereo applications rely on local algorithms in the disparity computation. These local algorithms generally suffer from matching ambiguities as it is difficult to find appropriate support for each pixel. Recent research shows that algorithms using adaptive cost aggregation approach greatly improve the quality of disparity map. Unfortunately, although these improvements are excellent, they are obtained at the expense of high computational. This article presents a hardware implementation for speeding up these methods. With hardware friendly approximation, we demonstrate the feasibility of implementing this expensive computational task on hardware to achieve real-time performance. The entire stereo vision system, includes rectification, stereo matching, and disparity refinement, is realized using a single field programmable gate array. The highly parallelized pipeline structure makes system be capable to achieve 51 frames per second for 640 × 480 stereo images. Finally, the success of accuracy improvement is demonstrated on the Middlebury dataset, as well as tests on real scene.

show abstract

Hardware-efficient belief propagation

Cited by 26 publications

References 34 publications

Detecting, segmenting and tracking unknown objects using multi-label MRF inference

Detecting, segmenting and tracking unknown objects using multi-label MRF inference

Software and Hardware Implementations of Stereo Matching

Real-time stereo vision system using adaptive weight cost aggregation approach

Contact Info

Product

Resources

About