Mar 10 (Wed): "Leveraging the Characteristics of Flow Between Homography-Warped Images for Applications in Graphics and 3D Vision," Abhishek Badki, ECE PhD Defense
A collection of images of a scene captured from different perspectives inform us about the scene's composition. Visualizing different perspectives of a scene in intuitive ways using simple techniques, like light field visualization, panorama image stitching, etc., allow us to maximize the perception and understanding of a scene. While visualizing a collection of images in intuitive ways allows humans to consume the image content in a much richer way, it can also provide better ways for computer algorithms to consume the novel visualizations of the captured images and reason about the 3D information in the scene. We explore these two threads by using a simple technique that uses plane-induced homography transformations, which are often used in image-based visualizations, to warp and compare images.
Plane-induced homography transformations allow us to warp and align multiple images with different perspectives of a scene to a common reference 3D plane. We use this simple image warping operation, combined with camera motion knowledge, to analyze the characteristics of flow between the warped images for the objects that lie on and off the 3D reference plane.
In part I of the thesis, we leverage the observed characteristics of the flow between homography-warped images to develop a computer graphics application that allows photographers to control different aspects of image composition in post-capture. This enables them to capture images that are not physically possible but artistically compelling. In part II, we also use it to develop a new framework that uses multiple visualizations for a pair of images to pose simple questions about the 3D scene to a neural-network and train it to provide binary answers. We use this framework to pose core 3D computer vision problems like depth estimation for static scenes and time-to-contact and optical flow estimation for dynamic scenes via binary classifications and show advantages over the existing approaches.
Abhishek Badki received his B.Tech in Electronics and Communication Engineering from the National Institute of Technology Trichy, India, in 2011, and his M.S. degree in Electrical and Computer Engineering from the University of California, Santa Barbara in 2015. He is currently a Ph.D. candidate in the ECE department at UCSB and is advised by Professor Pradeep Sen. His research interests lie at the intersection of machine learning, computer vision, and computer graphics.
Hosted by: Professor Pradeep Sen
Submitted by: Abhishek Badki <firstname.lastname@example.org>