Optical Flow Estimation in Synthetic Image Sequences Using Farneback Algorithm

Perspective motion corresponding to the real world motion between observer and scene can be characterized by a dense field corresponding to the interframe displacement of each pixel, known as optical flow. The analysis of optical flow has been an active research domain in the last couple of decades and tremendous progress has been attained by diverse approaches for its estimation. This paper demonstrates the implementation of Farneback method for optical flow determination. Various synthetic image sequences and their ground truth flow images from benchmark datasets are examined. Resultant flow fields, depicted as color-coded images, are compared with ground truth optical flow images by using Structural Similarity Index Measure whereas the flow vectors are compared with ground truth by estimating Angular Error and Endpoint Error.