Verification of bidirectional local based optical flow with bilateral filter on non-Gaussian noise contamination for video reconstruction

Abstract

In optical flow for motion approximation, the productive result in motion vector (MV) is an important issue for video reconstruction. Different in noisy conditions may cause the unreliable result in optical flow algorithms. A lot of robust optical flow algorithm was proposed for noise overthrown to increase the certain result under noisy conditions. This paper focuses on the efficiency of bidirectional local based optical flow with a bilateral filter on video reconstruction where the video sequence is contaminated with non-Gaussian noise (Poisson noise, Salt & Pepper noise, and Speckle noise). In our experimentation, the different set of video sequences in contamination with non-Gaussian noise are verified by the Peak Signal to Noise Ratio (PSNR) index.