Boundary ownership by lifting to 2.1D

Leichter, Ido; Lindenbaum, Michael

doi:10.1109/iccv.2009.5459208

Cited by 23 publications

(42 citation statements)

References 11 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…We simply attempt to replicate human behavior by training on human-annotated data. As previously pointed out [20,16], this means that figure/ground ordering does not necessarily correspond to depth or occlusion ordering. For example, humans may indicate strong figure/ground percepts due to markings on flat surfaces.…”

Section: Introductionmentioning

confidence: 72%

“…The most closely related work to ours is that of Ren et al [24] and Leichter and Lindenbaum [16], both of which focus on solving an easier problem than the one suggested here. Leichter and Lindenbaum take the human-drawn ground-truth segmentations [19] and human figure/ground annotations [10] of the BSDS images and learn a conditional random field (CRF) for assigning boundary ownership.…”

Section: Introductionmentioning

confidence: 96%

“…Historically, more attention has been paid to segmentation, though some important studies of figure/ground exist, focusing on contour and junction structure [13,11,25,32] or specific cues [10] such as convexity [21] or lower-region [29]. Recent work has revived interest on figure/ground discrimination [24,16] and the related problem of depth ordering [15,26].…”

Section: Introductionmentioning

confidence: 99%

“…Previous work starts from the assumption that figure/ground organization occurs after contours [24] or regions [16] have been obtained and designs algorithms that require an image segmentation as input. Hoiem et al [15] fix an initial oversegmentation and iterate region-merging and depth estimation steps.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Simultaneous Segmentation and Figure/Ground Organization Using Angular Embedding

Maire

2010

Computer Vision – ECCV 2010

View full text Add to dashboard Cite

Abstract. Image segmentation and figure/ground organization are fundamental steps in visual perception. This paper introduces an algorithm that couples these tasks together in a single grouping framework driven by low-level image cues. By encoding both affinity and ordering preferences in a common representation and solving an Angular Embedding problem, we allow segmentation cues to influence figure/ground assignment and figure/ground cues to influence segmentation. Results are comparable to state-of-the-art automatic image segmentation systems, while additionally providing a global figure/ground ordering on regions.

show abstract

Section: Introductionmentioning

confidence: 72%

Section: Introductionmentioning

confidence: 96%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Simultaneous Segmentation and Figure/Ground Organization Using Angular Embedding

Maire

2010

Computer Vision – ECCV 2010

View full text Add to dashboard Cite

show abstract

“…The problem to detect on occlusion boundaries which side is figure and which side is ground has been addressed by works in [14]- [16]. Similar to depth estimation systems, some of these approaches rely also on low level cues such as shapemes [14], convexity or parallelism [15]. The main drawback of these systems is that they do not provide closed partitions and only single contours are labeled.…”

Section: B Related Workmentioning

confidence: 99%

Monocular Depth Ordering Using T-Junctions and Convexity Occlusion Cues

Palou

Salembier

2013

IEEE Trans. on Image Process.

View full text Add to dashboard Cite

Abstract-This paper proposes a system that relates objects in an image using occlusion cues and arranges them according to depth. The system does not rely on a priori knowledge of the scene structure and focuses on detecting special points, such as T-junctions and highly convex contours, to infer the depth relationships between objects in the scene. The system makes extensive use of the binary partition tree as hierarchical region-based image representation jointly with a new approach for candidate T-junction estimation. Since some regions may not involve T-junctions, occlusion is also detected by examining convex shapes on region boundaries. Combining T-junctions and convexity leads to a system which only relies on low level depth cues and does not rely on semantic information. However, it shows a similar or better performance with the state-of-the-art while not assuming any type of scene.As an extension of the automatic depth ordering system, a semi-automatic approach is also proposed. If the user provides the depth order for a subset of regions in the image, the system is able to easily integrate this user information to the final depth order for the complete image. For some applications, user interaction can naturally be integrated, improving the quality of the automatically generated depth map.

show abstract

DOOBNet: Deep Object Occlusion Boundary Detection from an Image

Wang

et al. 2019

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Object occlusion boundary detection is a fundamental and crucial research problem in computer vision. Solving this problem is challenging as we encounter extreme boundary/non-boundary class imbalance during the training of an object occlusion boundary detector. In this paper, we propose to address this class imbalance by up-weighting the loss contribution of false negative and false positive examples with our novel Attention Loss function. We also propose a unified end-to-end multi-task deep object occlusion boundary detection network (DOOB-Net) by sharing convolutional features to simultaneously predict object boundary and occlusion orientation. DOOBNet adopts an encoderdecoder structure with skip connection in order to automatically learn multi-scale and multi-level features. We significantly surpass the stateof-the-art on the PIOD dataset (ODS F-score of .702) and the BSDS ownership dataset (ODS F-score of .555), as well as improving the detecting speed to as 0.037s per image on the PIOD dataset.

show abstract

Boundary ownership by lifting to 2.1D

Abstract: This paper addresses the "boundary ownership" problem, also known as the figure/ground

Cited by 23 publications

References 11 publications

Simultaneous Segmentation and Figure/Ground Organization Using Angular Embedding

Simultaneous Segmentation and Figure/Ground Organization Using Angular Embedding

Monocular Depth Ordering Using T-Junctions and Convexity Occlusion Cues

DOOBNet: Deep Object Occlusion Boundary Detection from an Image

Contact Info

Product

Resources

About