Verification of Video Reconstruction Using Bilateral-Reverse Directional Global Based Optical Flow over Non-Gaussian Noise
Darun Kesrarat and Vorapoj Patanavijit (Assumption University, Thailand)
This paper presents the verification of video reconstruction over non-Gaussian noise using Bilateral-reverse directional global based optical flow. Noise is a major problem in optical for motion approximation where the effective result in motion vector (MV) is a core matter in the optical flow. Various type of noise cause the uncertain determination in the optical flow technique. Many of robust optical flow techniques were invented to increase the certainty of optical under noisy conditions. In our experimentation, we verified several of robust optical flow techniques with the Bilateral-reverse directional technique. The contaminated video with non-Gaussian noises is used to verify the performance on several robust optical flow techniques by considering the result in the video reconstruction from the MV on each technique to prove the robustness of Bilateral-reverse directional technique on global based optical flow.
Journal: International Journal of Simulation: Systems, Science and Technology IJSSST V19
Published: Jun 30, 2018