Introducing In-Frame Shear Constraints for Monocular Motion Segmentation

In this thesis, the problem of motion segmentation is discussed. The aim of motion segmentation is to decompose a video into different objects that move through the sequence. In many computer vision pipelines, this is an important, middle step. It is essential in several applications like robotics, visual surveillance and traffic monitoring. While,there is already a vast amount of literature on the topic, the performance of all thus-far proposed algorithms are far behind human perception. This thesis starts of with a formal introduction to the problem. Then, it proceeds to explain the main approaches proposed to the problem, along with their advantages, and shortcomings. Finally, the proposed algorithm, that forms the keystone of this thesis, is introduced and fully-fleshed out, giving motivation for the structure and the various parts of the algorithm. In addition, the traditional comparison is given with the other-proposed state-of-the art algorithms. We do so, on the standard benchmark Hopkins-155 dataset, as well as a new dataset, compiled from video sequences from the publically available, KITTI dataset, the Versailles-Rond sequence taken from [] and several sequences taken around the IIIT Hyderabad campus. The sequences in the dataset, consist of video footage taken from a singlecamera mounted on the front of a car. The dataset is far more realistic and challenging than the Hopkins dataset, and provides a more rigorous assessment for both the proposed algorithm, as well as other state-of-the-art algorithms in motion segmentation. This dataset is hereby referred to as the On-Road dataset. On the Hopkins-155, our algorithm achieves near state-of-the-art performance, while performing substantially better on the On-Road dataset, showing that the proposed algorithm, has superior performance in realistic scenarios.