Enhancement of Ear Verification System Performance Using a New Hybrid Denoising Approach (ANFIS-FWS)

Image denoising is one of the important processes in the pre-processing of verification/identification systems. As a norm in verification system, Performance of feature extraction and approaches and verification depend son the input images quality. Since there is noise in the captured image via acquire/transfer sensors has direct effect on the quality of input image, insert a denoising step and enhancement before feature extraction system can be very effective, therefore using extracted features can prove actual or fake identity. This paper proposed a new method for increase the performance of an ear verification system, where At first, in a preprocessing step using hybrid denoising method, the noises are removed from ear images and then the next step these enhanced images (denoisy images) are used for verification system. Experimental results on this approach showed that this method well removed Gaussian noises from the ear dataset images and has acceptable impression on raising the verification accuracy.