2012
DOI: 10.1109/tpami.2011.271
|View full text |Cite
|
Sign up to set email alerts
|

Detachable Object Detection: Segmentation and Depth Ordering from Short-Baseline Video

Abstract: We describe an approach for segmenting a moving image into regions that correspond to surfaces in the scene that are partially surrounded by the medium. It integrates both appearance and motion statistics into a cost functional that is seeded with occluded regions and minimized efficiently by solving a linear programming problem. Where a short observation time is insufficient to determine whether the object is detachable, the results of the minimization can be used to seed a more costly optimization based on a… Show more

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
2
1
1
1

Citation Types

0
29
0

Year Published

2012
2012
2023
2023

Publication Types

Select...
3
3
2

Relationship

0
8

Authors

Journals

citations
Cited by 27 publications
(29 citation statements)
references
References 42 publications
0
29
0
Order By: Relevance
“…This approach only exploits the appearance cues, is applicable to a single view setting and more suitable in the context of image based retrieval applications, where images of scenes are well composed, containing little clutter. Our approach is motivated by work of [3], which explicitly reasons about evidence of occlusions boundaries extracted from optical flow and relative depth ordering cues. Also related to our work are several attempts to discover objects in urban scenes.…”
Section: Related Workmentioning
confidence: 99%
“…This approach only exploits the appearance cues, is applicable to a single view setting and more suitable in the context of image based retrieval applications, where images of scenes are well composed, containing little clutter. Our approach is motivated by work of [3], which explicitly reasons about evidence of occlusions boundaries extracted from optical flow and relative depth ordering cues. Also related to our work are several attempts to discover objects in urban scenes.…”
Section: Related Workmentioning
confidence: 99%
“…Southey and Little [8] provide another example of a live-video system, combining stereo vision with optical flow techniques to segment manipulable objects in video, and visual features to group these segments. Ayvaci and Soatto [9] use motion in video to find occlusion cues which are integrated to partition the image into depth layers. Sivic et al [10] do frame-to-frame tracking in video, and aggregate groups of points that move together to segment objects.…”
Section: Prior Workmentioning
confidence: 99%
“…Gestalt principles [33] provide grouping criteria: continuity, regularity, proximity, compactness, the last of which (figure/ground, or occlusion) is best informed by video. Occlusions have been used extensively for grouping [32,5,8,3]. A feature of [3] is that grouping is obtained via a linear program: local ordering constraints provided by occluder/occluded relations are integrated to globally partition the image domain into depth layers.…”
Section: Introductionmentioning
confidence: 99%