With the increase in technology, for the estimation and security biometric systems has created a various initiative. In this paper, we focused on various solutions that overcome the several kinds of issues faced by using hardware based fake fingerprint detection. In the existing scenario, to identify the spoofing of fake fingerprint, the system is trained with several kinds of images that represent the artificially moulded fingerprints also those images are compared with the real fingerprint images. This produces a low degree of complexity and gray scale levels in finger ridges and valleys. In our proposed scenario, we consider the features such as color, light transmission property fingerprint for features extraction. This efficient method discriminates the legitimate features from the artificially moulded fingerprints. Also, by using singular value decomposition method we can achieve more accurate detection of finger features. It is more robust to detect the fake features from the original features. From the experimental result, we can recognize the additional features such as color transmission and pressure features which are fake or real samples.