Results 1 
4 of
4
Salient object detection using concavity context
 In ICCV
, 2011
"... Convexity (concavity) is a bottomup cue to assign figureground relation in the perceptual organization [18]. It suggests that region on the convex side of a curved boundary tend to be figural. To explore the validity of this cue in the task of salient object detection, we segment the images in a ..."
Abstract

Cited by 15 (3 self)
 Add to MetaCart
(Show Context)
Convexity (concavity) is a bottomup cue to assign figureground relation in the perceptual organization [18]. It suggests that region on the convex side of a curved boundary tend to be figural. To explore the validity of this cue in the task of salient object detection, we segment the images in a test dataset into superpixels, and then locate the concave arcs and their bounding boxes along boundary of superpixels. Ecological statistics indicate that such bounding box contains salient object with a large probability. To utilize this spatial context information, i.e. concavity context, we follow the multiscale analysis of human visual perception and design a hierarchical model. The model yields an affinity graph over candidate superpixels, in which weights between vertices are determined by the summation of concavity context on different scales in the hierarchy. Finally a graphcut algorithm is performed to separate the salient and background objects. Evaluation on MSRA Salient Object Detection (SOD) dataset shows that concavity context is effective, and our approach provides improvement over stateoftheart featurebased algorithms. 1.
Measure of Circularity for Parts of Digital Boundaries and its Fast Computation
, 2010
"... ..."
(Show Context)
Fast recognition of a Digital Straight Line Subsegment: two algorithms of logarithmic time complexity
, 2013
"... Given a Digital Straight Line (DSL) of known characteristics (a, b, µ), we address the problem of computing the characteristics of any of its subsegments. We propose two new algorithms that use the fact that a digital straight segment (DSS) can be defined by its set of separating lines. The represen ..."
Abstract
 Add to MetaCart
(Show Context)
Given a Digital Straight Line (DSL) of known characteristics (a, b, µ), we address the problem of computing the characteristics of any of its subsegments. We propose two new algorithms that use the fact that a digital straight segment (DSS) can be defined by its set of separating lines. The representation of this set in the Z 2 space leads to a first algorithm of logarithmic time complexity. This algorithm precises and extends existing results for DSS recognition algorithms. The other algorithm uses the dual representation of the set of separating lines. It consists of a smart walk in the so called Farey Fan, which can be seen as the representation of all the possible sets of separating lines for DSSs. Indeed, we take profit of the fact that the Farey Fan of order n represents in a certain way all the digital segments of length n. The computation of the characteristics of a DSL subsegment is then equivalent to the localization of a point in the Farey Fan. Using fine arithmetical properties of the fan, we design a fast algorithm of theoretical complexity O(log(n)) where n is the length of the subsegment. Experiments show that our algorithms are also efficient in practice, with a comparison to the ones previoulsy proposed by Lachaud and Said [1]: in particular, the second one is much faster in the case of “small ” segments.
Measure of Circularity for Digital Curves and its Fast Computation
, 2008
"... This paper focuses on the design of an effective method that computes the measure of circularity of an open or closed digital curve. Thanks to its geometric interpretation, an algorithm that only uses classical tools of computational geometry is derived. Even if a sophisticated machinery coming from ..."
Abstract
 Add to MetaCart
(Show Context)
This paper focuses on the design of an effective method that computes the measure of circularity of an open or closed digital curve. Thanks to its geometric interpretation, an algorithm that only uses classical tools of computational geometry is derived. Even if a sophisticated machinery coming from linear programming can provide a linear time algorithm, its O(n log n) time complexity is better than many quadratic methods based on Voronoi diagrams. Moreover, this bound can be improve in the case of convex digital curves to reach linear time. 1