Combining Top-Down and Bottom-Up Segmentation

Borenstein, Eran; Sharon, E.; Ullman, Shimon

doi:10.1109/cvpr.2004.314

Cited by 204 publications

(172 citation statements)

References 15 publications

Supporting

Mentioning

166

Contrasting

Unclassified

Order By: Relevance

“…We have tested our algorithm on three widely used data sets: the extended Weizmann Horses [2,22], the ETHZ shapes [7] and the TU Darmstadt Database [16]. During the testing for Weizmann Horses, only 12 automatically selected horse silhouettes with one hand decomposed horse are used to learn the shape model.…”

Section: Resultsmentioning

confidence: 99%

“…In order to show the advantages of the proposed approach, we test our method on three widely used data sets, Weizmann horses [2], the ETHZ [6], and the cow dataset from the PASCAL Object Recognition Database Collection (TU Darmstadt Database [16]). Our results measured by bounding box intersection are comparable to state-of-the-art methods.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Weakly Supervised Shape Based Object Detection with Particle Filter

Yang

Latecki

2010

Computer Vision – ECCV 2010

View full text Add to dashboard Cite

Abstract. We describe an efficient approach to construct shape models composed of contour parts with partially-supervised learning. The proposed approach can easily transfer parts structure to different object classes as long as they have similar shape. The spatial layout between parts is described by a non-parametric density, which is more flexible and easier to learn than commonly used Gaussian or other parametric distributions. We express object detection as state estimation inference executed using a novel Particle Filters (PF) framework with static observations, which is quite different from previous PF methods. Although the underlying graph structure of our model is given by a fully connected graph, the proposed PF algorithm efficiently linearizes it by exploring the conditional dependencies of the nodes representing contour parts. Experimental results demonstrate that the proposed approach can not only yield very good detection results but also accurately locates contours of target objects in cluttered images.

show abstract

Section: Resultsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Weakly Supervised Shape Based Object Detection with Particle Filter

Yang

Latecki

2010

Computer Vision – ECCV 2010

View full text Add to dashboard Cite

show abstract

“…Fig. 6 shows a few more examples of image segmentation with the proposed method on the Weizemann dataset [14]. The upper row shows the original image and the lower row shows the segmentation results obtained after 4,4,3 iterations, correspondingly.…”

Section: Image Segmentationmentioning

confidence: 99%

“…In our experiments, we use several images from the Weizemann dataset consisting of 328 horse images [14] of gray scale. Each pixel of an image is viewed as a data point and the distance between two pixels is the difference of their gray value and the Euclidean distance between their coordinates.…”

Section: Image Segmentationmentioning

confidence: 99%

Distance Learning Based on Convex Clustering

Yang

Latecki

Gross

2009

Advances in Visual Computing

View full text Add to dashboard Cite

Abstract. Clustering has been among the most active research topics in machine learning and pattern recognition. Though recent approaches have delivered impressive results in a number of challenging clustering tasks, most of them did not solve two problems. First, most approaches need prior knowledge about the number of clusters which is not practical in many applications. Second, non-linear and elongated clusters cannot be clustered correctly. In this paper, a general framework is proposed to solve both problems by convex clustering based on learned distance. In the proposed framework, the data is transformed from elongated structures into compact ones by a novel distance learning algorithm. Then, a convex clustering algorithm is used to cluster the transformed data. Presented experimental results demonstrate successful solutions to both problems. In particular, the proposed approach is very suitable for superpixel generation, which are a common base for recent high level image segmentation algorithms.

show abstract

“…A variety of applications such as object recognition, image annotation, image coding and image indexing, utilize at some point a segmentation algorithm and their performance depends highly on the quality of the latter. Comparatively to the research efforts in automatic image and video segmentation [8], [18] and global [9], [14] or region-based [3], [13] image classification, still, human vision perception outperforms state-of-the-art computer algorithms. The main reason for this is that human vision is additionally based on high level a priori knowledge about the semantic meaning of the objects that compose the image and on contextual knowledge about their relationships.…”

Section: Introductionmentioning

confidence: 99%

A Context-Based Region Labeling Approach for Semantic Image Segmentation

Athanasiadis

Mylonas

Avrithis

2006

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Abstract. In this paper we present a framework for simultaneous image segmentation and region labeling leading to automatic image annotation. The proposed framework operates at semantic level using possible semantic labels to make decisions on handling image regions instead of visual features used traditionally. In order to stress its independence of a specific image segmentation approach we applied our idea on two region growing algorithms, i.e. watershed and recursive shortest spanning tree. Additionally we exploit the notion of visual context by employing fuzzy algebra and ontological taxonomic knowledge representation, incorporating in this way global information and improving region interpretation. In this process, semantic region growing labeling results are being re-adjusted appropriately, utilizing contextual knowledge in the form of domain-specific semantic concepts and relations. The performance of the overall methodology is demonstrated on a real-life still image dataset from the popular domains of beach holidays and motorsports.

show abstract

Combining Top-Down and Bottom-Up Segmentation

Abstract: In this work we show how to combine bottom-up and topdown approaches into a single figure-ground

Cited by 204 publications

References 15 publications

Weakly Supervised Shape Based Object Detection with Particle Filter

Weakly Supervised Shape Based Object Detection with Particle Filter

Distance Learning Based on Convex Clustering

A Context-Based Region Labeling Approach for Semantic Image Segmentation

Contact Info

Product

Resources

About