This document discusses techniques for detecting fake biometrics using image quality assessment. It begins with an introduction to biometric systems and the threats of fake biometric samples. It then describes using image quality measures to differentiate between real and fake fingerprint images. Twenty-five quality measures are extracted from each input image and used to train a support vector machine classifier. The trained SVM can then classify new input images as real or fake based on the quality scores. Experimental results show that wiener filtering improves the quality of fingerprint images for use in the classification system compared to Gaussian filtering. The system provides an effective software-based approach for fake biometric detection.