Distortions in fingerprint images arising due to the elasticity of fingerprints and pressure and movement of fingers during image capture lead to great difficulties in establishing a match between two images of a single fingerprint. In this paper, we propose a novel approach to detect and estimate distortion occurring in fingerprint videos. Our approach directly works on MPEG1 and MPEG-2 encoded fingerprint video bitstreams to estimate interfield flow without decompression and uses flow information to determine elastic distortion.