A new approach for fingerprint recognition based on mel frequency cepstral coefficients

This paper presents a new fingerprint identification method based on mel frequency cepstral coefficients (MFCCs). This method is based on the transformation of the fingerprint images into a 1-D signals and the extraction of the MFCCs from these signals. Feature matching is performed in the proposed method using neural networks. Wavelet transform of the 1-D fingerprint signals can be used for extraction of much more MFCCs to assist in the recognition process. Also wavelet denoising can be used to reduce noise. Experimental results show that the proposed method is robust in the presence of noise.