Fingerprint classification is a key issue inautomatic fingerprint identification systems. One of the maingoals is to reduce the item search time within the fingerprintdatabase without affecting the accuracy rate. In this paper, anovel technique, based on topological information, forefficient fingerprint classification is described. The proposedsystem is composed of two independent modules: the formermodule, based on Fuzzy C-Means, extracts the best set oftraining images; the latter module, based on Fuzzy C-Meansand Naive Bayes classifier, assigns a class to each processedfingerprint using only directional image information. Theproposed approach does not require any image enhancementphase. Experimental trials, conducted on a subset of the freedownloadable PolyU database, show a classification rate of91% over a 100 images test database using only 12 trainingexamples.
|Numero di pagine||7|
|Stato di pubblicazione||Published - 2014|
All Science Journal Classification (ASJC) codes
- Control and Systems Engineering
- Artificial Intelligence