Towards Feature Fusion - The Synthesis of Contour Sections Distinguishing Contours from Different Classes

Authors:
Dag Pechtel;Klaus-Dieter Kuhnert
Affiliations:
-;-
Venue:
DGCI '00 Proceedings of the 9th International Conference on Discrete Geometry for Computer Imagery
Year:
2000

Citing 8
Cited 0

2-D shape blending: an intrinsic solution to the vertex path problem

SIGGRAPH '93 Proceedings of the 20th annual conference on Computer graphics and interactive techniques
Parts of Visual Form: Computational Aspects

IEEE Transactions on Pattern Analysis and Machine Intelligence
Shapes, shocks, and deformations I: the components of two-dimensional shape and the reaction-diffusion space

International Journal of Computer Vision
A new method of estimating shape similarity

Pattern Recognition Letters
Shape Decomposition and Shape Similarity Measure

Mustererkennung 1998, 20. DAGM-Symposium
Discrete Approach to Curve Evolution

Mustererkennung 1998, 20. DAGM-Symposium
Automatic Generation of Significant and Local Feature Groups of Complex and Deformed Objects

ICIAP '99 Proceedings of the 10th International Conference on Image Analysis and Processing
Dynamic Programming

Dynamic Programming

Quantified Score

Hi-index	0.00

Visualization

Abstract

In real world problems, where the objects are in general complex and deformed, the automatic generation of local characteristics is necessary in order to distinguish different object classes. This paper presents an approach towards the automatic synthesis of significant local contour sections of closed, discrete, complex, and deformed 2D-object contours for the distinction of different classes. Neighboring contour points are determined and synthesized (feature fusion) into feature groups. Exclusively with the help of these feature groups the method distinguishes between different 2D-object contour classes of a certain domain. The basic idea is to get only the necessary information of a contour or a contour class for recognition.