Novel multimodal identification technique using Iris & Palmprint traits with various matching score level proportions using BTC of bit plane slices

Abstract

In multimodal biometric techniques are considered for fusion at different levels like Score level, Feature level and Decision level. Here the Iris and Palmprint traits considered with score level fusion to get proposed multimodal identification technique using Block Truncation Coding (BTC) applied on the individual biplanes of iris and palmprint images. Use of Block Truncation Code makes the feature extraction independent of size of iris and palmprint images. The experimentation done using test bed with 60 pairs of iris and palmprint images for 10 persons. Experimentation results have indicated that BTC level 1 performs better than BTC level 2 in all biplanes for proposed multimodal biometric identification technique. For the score level fusion of features of iris and palm traits various proportions are used. The higher proportion of Palmprint gives better identification. The proposed multimodal identification techniques with score level Iris: Palmprint fusion with 1:4 proportions has given best genuine acceptance rate with BTC level 1.