1,720,964 research outputs found
Dominant points detection on digital curves: A comparison between optimal and exact approaches
No full textIn this work we address the problem of closed digital curves polygonal approximation by locating a set of relevant points having high curvature, the so-called dominant points. This set of feature points plays a dominant role in shape perception by humans and contains almost complete information of a given contour. There are several methods to extract dominant points based on different approaches; we look over two heuristic techniques, based on Ant Colony Optimization (ACO) and based on Genetic Algorithm (GAs), and an original method based on Dominant Points Iterative Localization (DP1L). We compare the three algorithms by evaluating the approximation error and testing their affine transformations invariance
A new iterative approach for dominant points extraction in planar curves
No full textIn this paper the problem of dominant point detection on digital curves is addressed. Based on an initial set of curvature points, our approach adds iteratively significant points by looking for the higher curvature contour points. The process continues until all the sums of the distances of contour points in the arcs subtended to the chord between two next dominant points is less then a predefined threshold. A final refinement process adjusts the position of located dominant points by a minimum integral square error criterion. We test our method by comparing its performance with other well known dominant point extraction techniques succesfully. In the last section some examples of polygonal approximation are shown
Shape matching by curve modelling and alignment
Full text linkAutomatic information retrieval in the eld of shape recognition has been widely covered by many
research elds. Various techniques have been developed using different approaches such as intensity-based, modelbased
and shape-based methods. Whichever is the way to represent the objects in images, a recognition method
should be robust in the presence of scale change, translation and rotation. In this paper we present a new recognition
method based on a curve alignment technique, for planar image contours. The method consists of various phases
including extracting outlines of images, detecting signicant points and aligning curves. The dominant points can
be manually or automatically detected. The matching phase uses the idea of calculating the overlapping indices
between shapes as similarity measures. To evaluate the effectiveness of the algorithm, two databases of 216 and
99 images have been used. A performance analysis and comparison is provided by precision-recall curves
A Comparison of 2-D Moment-Based Description Techniques
No full textMoment invariants are properties of connected regions in binary images that are invariant to translation, rotation and scale, They are useful because they define a simply calculated set of region properties that can be used for shape classification and part recognition, Orthogonal moment invariants allow for accurate reconstruction of the described shape. Generic Fourier Descriptors yield spectral features and have better retrieval performance due to multi-resolution analysis in both radial and circular directions of the shape. In this paper we first compare various moment-based shape description techniques then we propose a method that, after a previous image partition into classes by morphological features, associates the appropriate technique with each class, i.e. the technique that better recognizes the images of that class, The results clearly demonstrate the effectiveness of this new method regard to described techniques
A new algorithm for polygonal approximation based on ant colony optimization
No full textIn shape analysis a crucial step consists in extracting meaningful features from digital curves. Dominant points are those points with curvature extreme on the curve that can suitably describe the curve both for visual perception and for recognition. In this paper we present a novel method that combines the dominant point detection and the ant colony
optimization search. The excellent results have been compared both to works using an optimal search approach and to works based on exact approximation strategy
Moment-based Techniques for Image Retrieval
No full textIn this paper we analyze some shape-based image retrieval methods which use different types of geometric and algebraic moments and Fourier descriptors. Moments have been widely used in pattern recognition applications to describe the geometrical characteristics of different objects. They provide fundamental geometric properties (e.g. area, centroid, moment of inertia, etc..). We consider various description techniques: Hu, Flusser and Taubin invariants, Legendre and Zernike moments, Generic Fourier Descriptors (GFD). The set of absolute orthogonal (i.e. rotation) moment invariants defined by Hu can be used for scale, position, and rotation invariant pattern identification. Flusser' s complete set of invariants appears as a particular case, with invariance only to rotation. The Taubin's affine moment invariants introduce the concept of covariant matrix. Legendre moments are based on orthogonal Legendre polynomials and are not invariant under image rotation. Zernike moments consist of a set of complex polynomials that form a complete orthogonal set over the interior of the unit circle. GFDs are derived by applying a modified polar Fourier transform on shape image. We have applied the retrieval methods on a collection of images chosen from MPEG7 database. The image retrieval performance of each method is described by the precision-recall graph. In the paper we propose a novel approach that combines the described techniques after a coarse partitioning of the image dataset by their morphological features
A new genetic algorithm for polygonal approximation
No full textIn this chapter, the problem of approximating a closed digital curve with a simplified representation by a set of feature points containing almost complete information of the contour, i.e., dominant points, is addressed. We adopt an approach based on genetic algorithms (GAs) since they use parallel search and have good performance in solving optimization problems. The chromosome coincides with an approximating polygon and is represented by a binary string. Each bit, called gene, represents a curve point where dominant points have 1-value. The proposed algorithm enhances the selection and mutation phase avoiding the premature convergence issue. Our method is compared to other similar approaches and its efficiency is clearly demonstrated by experimental results giving a better approximation by lowering the error norm with respect to the original curves
- …
