Search for Similar Profiles through Pattern Matching Algorithms:
Alain Michalak, Compes France, Brouchy, France,
Effectiveness and Advantages for Extrusion Companies
Marco Azzalin, Stefano Bellotti, Vincenzo Giacomelli,
Compes SpA, Rodengo Saiano, Italy.
Track: MANAGEMENT ISSUES - Operations
ABSTRACT—Pattern-matching algorithms are widely used for automatic fingerprint identification and in computer vision.
Compes, one of the leading extrusion die and tool supplier companies, developed the program
SENDRA that uses adapted algorithms to effectively work with aluminum profile shapes. The program
allows extruders to search for similar profiles in a fast and easy way.
To make this method useful, it is essential to link the search results with additional data regarding the
press parameters and the productivity of the profile. In this way the comparison between similar profiles
can be used to optimize press parameters and precisely define quotations for new profiles, based on
previous similar profiles’ productivity.
In this paper, we will describe the search algorithm, evaluate its effectiveness and analyze real usage
scenarios in some extruders companies.
© Extrusion Technology for Aluminum Profiles Foundation (ET Foundation). All rights reserved. No part of The Proceedings may be reproduced in any form without the express written permission of the ET Foundation.