Abstract-We describe a geometric-flow-based algorithm for computing a dense oversegmentation of an image, often referred to as superpixels. It produces segments that, on one hand, respect local image boundaries, while, on the other hand, limiting undersegmentation through a compactness constraint. It is very fast, with complexity that is approximately linear in image size, and can be applied to megapixel sized images with high superpixel densities in a matter of minutes. We show qualitative demonstrations of high-quality results on several complex images. The Berkeley database is used to quantitatively compare its performance to a number of oversegmentation algorithms, showing that it yields less undersegmentation than algorithms that lack a compactness constraint while offering a significant speedup over N-cuts, which does enforce compactness.

OBJECTIVE: To develop an automatic system that grades the severity of facial signs through 'selfies' pictures taken by women of different ages and ethnics. METHODS: 1140 women from three ethnics (African-American, Asian, Caucasian), of different ages (18-80 years old), took 'selfies' by high resolution smartphones cameras under different conditions of lighting or facial expressions. A dedicated software, was developed, based on a Convolutional Neural Network (CNN) that integrates training data from referential Skin Aging Atlases. The latter allows to an immediate quantification of the severity of nine facial signs according to the ethnicity declared by the subject. These automatic grading were confronted to those assessed by 12 trained experts and dermatologists either on 'selfies' pictures or in live conditions on a smaller cohort of women. RESULTS: The system appears weakly influenced by lighting conditions or facial expressions (coefficients of variations ranging 10-13% for most signs) and leads to global agreements with experts' assessments, even showing a better reproducibility on some facial signs. CONCLUSION: This automatic scoring system, still in development, seems offering a new quantitative approach in the quantified description of facial signs, independent from human vision, in many applications, being individual, cosmetic oriented or dermatological with regard to the follow-up of medical anti-ageing corrective strategies. a des valeurs comparables de celles des experts, voire même de meilleure reproductibilit e dans certains cas. CONCLUSION: Ce syst eme de scorage automatique, encore en d eveloppement, semble offrir une nouvelle approche dans la description quantitative de signes du visage, ind ependante de l'oeil humain, dans de nombreuses applications, comme la personnalisation, a vis ee cosm etique ou dermatologique, dans le suivi de certaines strat egies m edicales de l'antivieillissement cutan e.

Skeletonization algorithms typically decompose an object's silhouette into a set of symmetric parts, offering a powerful representation for shape categorization. However, having access to an object's silhouette assumes correct figure-ground segmentation, leading to a disconnect with the mainstream categorization community, which attempts to recognize objects from cluttered images. In this paper, we present a novel approach to recovering and grouping the symmetric parts of an object from a cluttered scene. We begin by using a multiresolution superpixel segmentation to generate medial point hypotheses, and use a learned affinity function to perceptually group nearby medial points likely to belong to the same medial branch. In the next stage, we learn higher granularity affinity functions to group the resulting medial branches likely to belong to the same object. The resulting framework yields a skeletal approximation that's free of many of the instabilities plaguing traditional skeletons. More importantly, it doesn't require a closed contour, enabling the application of skeleton-based categorization systems to more realistic imagery.

Abstract. Detecting contour closure, i.e., finding a cycle of disconnected contour fragments that separates an object from its background, is an important problem in perceptual grouping. Searching the entire space of possible groupings is intractable, and previous approaches have adopted powerful perceptual grouping heuristics, such as proximity and co-curvilinearity, to manage the search. We introduce a new formulation of the problem, by transforming the problem of finding cycles of contour fragments to finding subsets of superpixels whose collective boundary has strong edge support in the image. Our cost function, a ratio of a novel learned boundary gap measure to area, promotes spatially coherent sets of superpixels. Moreover, its properties support a global optimization procedure using parametric maxflow. We evaluate our framework by comparing it to two leading contour closure approaches, and find that it yields improved performance.

Skeletonization algorithms typically decompose an object's silhouette into a set of symmetric parts, offering a powerful representation for shape categorization. However, having access to an object's silhouette assumes correct figure-ground segmentation, leading to a disconnect with the mainstream categorization community, which attempts to recognize objects from cluttered images. In this paper, we present a novel approach to recovering and grouping the symmetric parts of an object from a cluttered scene. We begin by using a multiresolution superpixel segmentation to generate medial point hypotheses, and use a learned affinity function to perceptually group nearby medial points likely to belong to the same medial branch. In the next stage, we learn higher granularity affinity functions to group the resulting medial branches likely to belong to the same object. The resulting framework yields a skeletal approximation that's free of many of the instabilities plaguing traditional skeletons. More importantly, it doesn't require a closed contour, enabling the application of skeleton-based categorization systems to more realistic imagery.

No abstract

No abstract

The results demonstrated the feasibility of utilizing a learning-based approach for automatic selection of beam angles in thoracic IMRT planning. The proposed method may assist in reducing the manual planning workload, while sustaining plan quality.