Laboratoire d'Informatique Gaspard Monge

Current content-based image retrieval techniques can typically perform efficient and effective searches on heterogeneous image databases. This contribution deals with an approach based on the integration of color and texture description which is applied to a very homogeneous database: a blood image bank. The content of images is very similar and therefore it becomes imperative to use very precise descriptors: the color is described by classical color distributions (histograms) and for the texture, we introduce the morphological color size distributions. The similarity is measured by computing distance metrics between histograms. In order to increase the accuracy of retrieval, the results of color-based and texture-based retrieval are integrated by combining the associated dissimilarity values. The effects of different integration methods on classification performance are explained by means of experimental tests in a database of 123 cell images (leukocyte color images). After learning processing, where different feature selection and classifier definition alternatives are tested, a definitive integrated approach is proposed (precision ¢£¤££¥ ).

This paper deals with the theory and applications of spatially-variant discrete mathematical morphology. We review and formalize the definition of spatially variant dilation/erosion and opening/closing for binary and gray-level images using exclusively the structuring function, without resorting to complement. This theoretical framework allows to build morphological operators whose structuring elements can locally adapt their shape and orientation across the dominant direction of the structures in the image. The shape and orientation of the structuring element at each pixel are extracted from the image under study: the orientation is given by means of a diffusion process of the average square gradient field, which regularizes and extends the orientation information from the edges of the objects to the homogeneous areas of the image; and the shape of the orientated structuring elements can be linear or it can be given by the distance to relevant edges of the objects. The proposed filters are used on binary and gray-level images for enhancement of anisotropic features such as coherent, flow-like structures. Results of spatially-variant erosions/dilations and openings/closings-based filters prove the validity of this theoretical sound and novel approach.

Wkh pruskrorjlfdo0d!qh remhfw ghirupdwlrq Pduflq Lzdqrzvnl dqg Mhdq Vhuud Fhqwuh gh Pruskrorjlh Pdwkìpdwltxh/ Hfroh ghv Plqhv gh Sdulv/ 68/ uxh Vdlqw0Krqruì/ ::638 Irqwdlqheohdx +IUDQFH, LVPP 5333 SdorDowr/ Mxqh 5333 Devwudfw Wkh frpelqdwlrq ri pruskrorjlfdo lqwhusrodwlrq dqg d!qh wudqviru0 pdwlrq lv suhvhqwhg1 Wkh sursrvhg dssurdfk xqlwhv wkh dgydqwdjhv ri erwk phwkrgv= wkh glvsodfhphqw lv shuiruphg e| xvlqj d!qh wudqvirupd0 wlrq/ dqg wkh vkdsh ghirupdwlrq e| pruskrorjlfdo lqwhusrodwlrq1 Lw doorzv wkh wudqvirupdwlrq ri rqh elqdu| vhw lqwr dqrwkhu lq vhpl0dxwrpdwlf ru ixoo|0dxwrpdwlf zd|1 dwkhpdwlfdo pruskrorj|/ lqwhusrodwlrq/ d!qh wudqvirupdwlrq/ remhfwv ghirupdwlrq © ©

Hierarchical segmentation is a multi-scale analysis of an image and provides a series of simplifying nested partitions. Such a hierarchy is rarely an end by itself and requires external criteria or heuristics to solve problems of image segmentation, texture extraction and semantic image labelling. In this theoretical paper we introduce a novel framework: hierarchical cuts, to formulate optimization problems on hierarchies of segmentations. Second we provide the three important notions of h-increasing, singular, and scale increasing energies, necessary to solve the global combinatorial optimization problem of partition selection and which results in linear time dynamic programs. Common families of such energies are summarized, and also a method to generate new ones is described. Finally we demonstrate the application of this framework on problems of image segmentation and texture enhancement.