Relational matching is a method for finding the bestcorrespondences betweenstructural descriptions. It iswidely used in computer vision for the recognition andlocation of objects in digital images. For this purpose, thedigital images and the object models are represented bystructural descriptions. The matching algorithm then has todetermine which image elements and object model partscorrespond.This book is the result of abasic study of relationalmatching. The book focuses particularly on the evaluation ofcorrespondences. In order to find the best match, one needsa measure to evaluate the quality of a match. The authorreviews the evaluation measures that have been suggestedover the past few decades and presents a new measure basedon information theory. The resulting theorycombinesmatching strategies, information theory, and tree searchmethods. For the benefit of the reader, comprehensiveintroductions are given to all these topics.
Computer vision and matching.- A classification of matching methods.- Formal description of relational matching.- Problem definition and contributions of the thesis.- Information theory:Selected Topics.- Evaluation of mappings between relational descriptions.- Tree search methods and heuristics.- Relational image and model description.- Evaluation functions for object location.- Strategy and performance of the tree search for object location.- Summary and discussion.
Series: Lecture Notes in Computer Science
Number Of Pages: 18
Published: 10th September 1992
Publisher: Springer-Verlag Berlin and Heidelberg Gmbh & Co. Kg
Country of Publication: DE
Dimensions (cm): 23.39 x 15.6
Weight (kg): 0.3