Piecewise Flat Embedding for Image Segmentation

Yu, Yizhou; Fang, Chaowei; Liao, Zuwei

doi:10.1109/iccv.2015.161

Cited by 24 publications

(58 citation statements)

References 22 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…It is also worth pointing out that our hierarchical segmentation framework can be used in combination with other features that can better guide the boundary classification. For example, using the most recent piecewise flat embedding (PFE) [24], we expect the results to be further improved in a manner similar to the results from “MCG” to “PFE-MCG” on BSDS500 in Table IV.…”

Section: Methodsmentioning

confidence: 69%

“…In this section, we compare our proposed iterative hierarchical merge tree method (CCM + ensemble boundary classifier + iteration, under name “HMT”) with various other state-of-the-art region segmentation methods and benchmarks [1], [25], [26], [23], [49], [22], [24] in very recent years on the public data sets. The results are shown in Table IV.…”

Section: Methodsmentioning

confidence: 99%

See 1 more Smart Citation

Image Segmentation Using Hierarchical Merge Tree

Liu

Seyedhosseini

Taşdizen

2016

IEEE Trans. on Image Process.

View full text Add to dashboard Cite

This paper investigates one of the most fundamental computer vision problems: image segmentation. We propose a supervised hierarchical approach to object-independent image segmentation. Starting with over-segmenting superpixels, we use a tree structure to represent the hierarchy of region merging, by which we reduce the problem of segmenting image regions to finding a set of label assignment to tree nodes. We formulate the tree structure as a constrained conditional model to associate region merging with likelihoods predicted using an ensemble boundary classifier. Final segmentations can then be inferred by finding globally optimal solutions to the model efficiently. We also present an iterative training and testing algorithm that generates various tree structures and combines them to emphasize accurate boundaries by segmentation accumulation. Experiment results and comparisons with other recent methods on six public data sets demonstrate that our approach achieves state-of-the-art region accuracy and is competitive in image segmentation without semantic priors.

show abstract

Section: Methodsmentioning

confidence: 69%

Section: Methodsmentioning

confidence: 99%

Image Segmentation Using Hierarchical Merge Tree

Liu

Seyedhosseini

Taşdizen

2016

IEEE Trans. on Image Process.

View full text Add to dashboard Cite

show abstract

“…of PFE+MCG [33]. It is noteworthy that our method and theirs are complementary, and the combination of the two may yield even better results.…”

Section: Comparison To Other Methodsmentioning

confidence: 84%

Scale-Aware Alignment of Hierarchical Image Segmentation

Chen

Dai

Pont-Tuset

et al. 2016

2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)

View full text Add to dashboard Cite

“…is either 'w' or 'o', denoting whether residual-based channel weighting is turned on or not; [size] is the width of neighborhood window(default value is 11); and [norm] is either '1' or 'p', '1' for the L 1,1 -regularized objective and 'p' for the L 1,p -regularized (p = 1) objective. For example, goPFE 1 , the same version as in [23], meaning PFE is computed using the L 1,1 -regularized objective in (8) and GMM based initialization without channel weighting. And swPFE p means PFE is computed using the L 1,p -regularized objective in (3), residual-based channel weighting and the WSC initialization.…”

Section: Metricsmentioning

confidence: 99%

Piecewise Flat Embedding for Image Segmentation

Fang

Liao

2019

IEEE Trans. Pattern Anal. Mach. Intell.

Self Cite

View full text Add to dashboard Cite

We propose a new multi-dimensional nonlinear embedding -- Piecewise Flat Embedding (PFE) -- for image segmentation. Based on the theory of sparse signal recovery, piecewise flat embedding with diverse channels attempts to recover a piecewise constant image representation with sparse region boundaries and sparse cluster value scattering. The resultant piecewise flat embedding exhibits interesting properties such as suppressing slowly varying signals, and offers an image representation with higher region identifiability which is desirable for image segmentation or high-level semantic analysis tasks. We formulate our embedding as a variant of the Laplacian Eigenmap embedding with an - regularization term to promote sparse solutions. First, we devise a two-stage numerical algorithm based on Bregman iterations to compute L1,1-regularized piecewise flat embeddings. We further generalize this algorithm through iterative reweighting to solve the general L1,p-regularized problem. To demonstrate its efficacy, we integrate PFE into two existing image segmentation frameworks, segmentation based on clustering and hierarchical segmentation based on contour detection. Experiments on four major benchmark datasets, BSDS500, MSRC, Stanford Background Dataset, and PASCAL Context, show that segmentation algorithms incorporating our embedding achieve significantly improved results.

show abstract

Piecewise Flat Embedding for Image Segmentation

Cited by 24 publications

References 22 publications

Image Segmentation Using Hierarchical Merge Tree

Image Segmentation Using Hierarchical Merge Tree

Scale-Aware Alignment of Hierarchical Image Segmentation

Piecewise Flat Embedding for Image Segmentation

Contact Info

Product

Resources

About