PROJECT TITLE :

Synthetic Video Sequence for Dynamic Scenes - 2017

ABSTRACT:

New approaches in image and video processing fields want a artificial dataset for testing the performance of their algorithms in order to improve their outcomes. Many of these approaches rely on a fastened lighting assumption. In the $64000 world, video surveillance may be exposed to different illumination or lighting conditions. During this paper, an algorithm for generating a dynamic blurry video dataset using Gaussian filter and Random Variety Generator is proposed to simulate the real world scenes beneath different illumination or lighting changes. The results showed that the proposed methodology can be used to generate blurred video dataset from the original video dataset beneath completely different illumination conditions.