Local matting based on sample-pair propagation and iterative refinement

He, Bin; Wang, Guijin; Ruan, Zhiwei; Yin, Xuanwu; Pei, Xiaokang; Lin, Xiao

doi:10.1109/icip.2012.6466851

Cited by 6 publications

(26 citation statements)

References 10 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To overcome problems listed above, we provided a local matting method which refined sample-pairs in a propagation mode and optimized the alpha matte with the kalman filter in our previous work [17]. This paper further expands the original idea and proposes an algorithm based on sample-pair refinement and local optimization.…”

Section: Introductionmentioning

confidence: 99%

“…Gastal [15] selected the most confident sample-pairs from ones in the neighborhood to refine the current sample-pair. In contrast, He [17] propagated sample-pairs among neighboring pixels. However, they neglected the fact that the foreground/background component of low confidence sample-pair may be useful.…”

Section: Related Work and Motivationmentioning

confidence: 99%

“…1) Sample-pair Initialization: Recent work [13]- [17] gathered multiple pixels in definite foreground/background regions as samples to initialize the alpha matte, named as color-sampling. Wang [13] collected dense samples on the basis of the minimal Euclidean distance.…”

Section: Related Work and Motivationmentioning

confidence: 99%

“…To achieve a semantically meaningful alpha matte [1], almost all matting approaches start by dividing the input image into 3 regions: definite foreground, definite background and the unknown (named as the trimap). Related matting methods can be categorized into sampling-based [2]- [4], affinity-based [5]- [11] and combinations of the two [12]- [17]. The combination-based one utilizes the relevance of color values between the foreground/background samples and current pixel, while the affinity of the alpha matte is guaranteed.…”

Section: Introductionmentioning

confidence: 99%

“…The combination-based one utilizes the relevance of color values between the foreground/background samples and current pixel, while the affinity of the alpha matte is guaranteed. Consequently, recent methods fall into the combination-based category [12]- [17]. Wang [13] firstly exploited the color-sampling strategy to estimate alpha values, followed by the random walk optimization [18] to smooth matting results.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

High-Accuracy and Quick Matting Based on Sample-Pair Refinement and Local Optimization

Wang

Shi

et al. 2013

IEICE Trans. Inf. & Syst.

Self Cite

View full text Add to dashboard Cite

SUMMARYBased on sample-pair refinement and local optimization, this paper proposes a high-accuracy and quick matting algorithm. First, in order to gather foreground/background samples effectively, we shoot rays in hybrid (gradient and uniform) directions. This strategy utilizes the prior knowledge to adjust the directions for effective searching. Second, we refine sample-pairs of pixels by taking into account neighbors'. Both high confidence sample-pairs and usable foreground/background components are utilized and thus more accurate and smoother matting results are achieved. Third, to reduce the computational cost of sample-pair selection in coarse matting, this paper proposes an adaptive sample clustering approach. Most redundant samples are eliminated adaptively, where the computational cost decreases significantly. Finally, we convert fine matting into a de-noising problem, which is optimized by minimizing the observation and state errors iteratively and locally. This leads to less space and time complexity compared with global optimization. Experiments demonstrate that we outperform other state-of-the-art methods in local matting both on accuracy and efficiency.

show abstract

Section: Introductionmentioning

confidence: 99%

Section: Related Work and Motivationmentioning

confidence: 99%