Dor Verbin

I am a computer science Ph.D. candidate at the Harvard School of Engineering and Applied Sciences, working with Todd Zickler. I am also currently a student researcher at Google Research, working with Jon Barron, Pratul Srinivasan, and Ben Mildenhall.

Research Interests:

I am interested in computer vision algorithms that are informed by geometry and physics. By incorporating geometric and physical constraints into differentiable architectures for visual inference, I aim to improve the efficiency and generalizability of artificial vision systems.

Projects:

	Field of Junctions: Extracting Boundary Structure at Low SNR Dor Verbin, Todd Zickler ICCV, 2021   By modeling each patch in an image as a generalized junction, our model uses concurrencies between different boundary elements such as junctions, corners, and edges, and manages to extract boundary structure from extremely noisy images where previous methods fail. [project page] [paper] [video] [code]
	Toward a Universal Model for Shape from Texture Dor Verbin, Todd Zickler CVPR, 2020   We formulate the shape from texture problem as a 3-player game. This game simultaneously estimates the underlying flat texture and object shape, and it succeeds for a large variety of texture types. [project page] [paper] [video] [supplement] [code and data]
	Unique Geometry and Texture from Corresponding Image Patches Dor Verbin, Steven J. Gortler, Todd Zickler TPAMI, 2021   We present a simple condition for the uniqueness of a solution to the shape from texture problem. We show that in the general case four views of a cyclostationary texture satisfy this condition and are therefore sufficient to uniquely determine shape. [preprint] [paper]