Classification and coding of contour information

1985

A generalised coding scheme is proposed for two-tone image contours. The basic idea is to detect digital line segments on the contour and code them using fixed length or variable length codewords. It can be shown that the conventional contour run length coding using 8 or 4 direction Freeman's chain code is a special case of the present scheme. The data compressibility is tested on several closed contours and their noisy degraded versions. The effect of gradual increases of noise level on the relative compressibility with respect to the conventional method is also studied, Extension of the work has also been proposed. +~

IEEE Transactions on Image Processing, 2017

1998

Abstract We propose a new method for the encoding of label images (also known as segmentation maps or alpha planes) that are often used to identify object location in region-based image and video coders. The method is contour-based and lossless with a contour model composed of two parts: a contour graph describing the topology of the contour network and a directional chain code to deal with the geometric part of the label image (internal contour points).

2003

With the rapid growth of computing power, many concepts and tools of image analysis are becoming more and more popular in other data processing fields, such as image and video compression. Image segmentation, in particular, has a central role in the object-based video coding standard MPEG-4, as well as in various region-based coding schemes used for remote-sensing imagery. A region-based image description, however, is only useful if it has a limited representation cost, which calls for accurate and efficient tools for the description of region boundaries. A very promising approach relies on the extended boundary concept, first discussed in [6] and [7] and later used by Liow [5] to develop a contour tracing algorithm. In this work, we extend Liow's algorithm and introduce the corresponding reconstruction technique needed for coding purposes. In addition, we define an algebraic semi-group structure that allows us to formally prove the algorithm properties, to extend it to other boundary definitions, and to introduce a fast contour tracing algorithm which only requires a raster scan of the image.

The tasks of image analysis are one of the fundamental problems of state-of-the-art image processing. The first contour decomposition algorithm into sets of finite and closed contours is presented in this paper. It is based on the representation of the vertex as a salient feature of an image object and developed by the author determination of a binary image pixel as a candidate for the vertex and closed contour detection algorithms. This algorithm puts into practice the developed by the author mathematical models of a contour and a particular image object on the computer. Note that a contour is a connected set. It is proposed that the particular image object model and contour decomposition task should be considered as the basis for a modern contour analysis. The acceptability of the proposed algorithm is shown. The solution, which allows us to perform detection of a particular image object, is obtained.

Video coding has been under intense scrutiny during the last years. The published international standards rely on low-level vision concepts, thus being first-generation. Recently standardization started in second-generation video coding, supported on mid-level vision concepts such as objects. This thesis presents new architectures for second-generation video codecs and some of the required analysis and coding tools. The graph theoretic foundations of image analysis are presented and algorithms for generalized shortest spanning tree problems are proposed. In this light, it is shown that basic versions of several region-oriented segmentation algorithms address the same problem. Globalization of information is studied and shown to confer different properties to these algorithms, and to transform region merging in recursive shortest spanning tree segmentation (RSST). RSST algorithms attempting to minimize global approximation error and using affine region models are shown to be very effective. A knowledge-based segmentation algorithm for mobile videotelephony is proposed. A new camera movement estimation algorithm is developed which is effective for image stabilization and scene cut detection. A camera movement compensation technique for first-generation codecs is also proposed. A systematization of partition types and representations is performed with which partition coding tools are overviewed. A fast approximate closed cubic spline algorithm is developed with applications in partition coding.

2001

This paper provides an algorithm for partitioning grayscale images into disjoint regions of coherent brightness and texture. Natural images contain both textured and untextured regions, so the cues of contour and texture differences are exploited simultaneously. Contours are treated in the intervening contour framework, while texture is analyzed using textons.

Information Sciences, 1996

This paper introduces a new operator for the calculation of the brightness gradient (derived from the Compass gradient) and a new methodology based on a fuzzy algorithm for extraction of the contour points of objects in a real image. A bidimensional geometrical representation of the contours thus extracted is then given in the form of arcs and segments in order to exploit this simplified description as an index for a search in a database of 3-D primitives. The procedure described gives good results, even in the presence of juxtaposed surfaces of the same nature and color, or with contours which do not contrast very well. 93, 159-185 (1996)

Image and Vision Computing, 1987

Edge-based binocular correspondence produces a sparse disparity map, available information being distributed along space curves which project to matched image edges. To become useful these contours must be parsed into describable sections. We present a novel view of the segmentation/description process and describe an effective algorithm based on our model.

In this thesis it is explored how far a computer can perform the task of preserving the content of an image while removing totally or partially its texture information. This problem is studied in two application fields: contour detection and artistic imaging. Concerning contour detection, the state of the art is first reviewed. Several classes of local algorithms are considered, which are mainly based on differential analysis, statistical approaches, phase congruency, rank order filters, and combinations thereof. More sophisticated global methods based on the computation of contour saliency, perceptual grouping, relaxation labeling and active contours are considered too. Then, algorithms for contour detector are developed, which are inspired by several aspects of both low and high level of the human visual system. In particular, a biologically motivated multiresolution contour detector with Bayesian denoising and surround inhibition is proposed, which takes into account several low ...