Vector graphic and image modeling research earns $499K NSF grant

 

Professor Hong Qin from the Department of Computer Science (CS) will use NSF funding to research novel methods for vectorization of arbitrary natural images and its applications.

Vector graphics are an important part of web design and image presentation. Affecting anything from image resolution, network bandwidth and animation, vector graphics are comprised of paths and mathematical commands instead of pixels which are found in your standard jpg images. Vector graphics offer a compact solution without sacrificing image representation.

While attempting to advance image vectorization, Qin’s research will delve into partial differential equations or PDEs and how they connect with Green’s functions and harmonic B-splines- something that has never been closely examined in the context of vector graphics.

According to Qin, “At the core of this project's theoretical foundation are PDEs and their meshless closed-form solvers based on fundamental solutions. This project will explore a novel image vectorization modeling scheme: Poisson Vector graphics (PVG), which computes complex color gradients via a sparse set of geometric primitives and color constraints. The novel representation is expected to outperform the conventional diffusion curve based and gradient mesh based representations.”

The goal of the project entitled Novel Method for Vectorization of Arbitrary Natural Images and Its Applications, is to advance vector graphics and image modeling in both practice and theory. The result will benefit not only the field of computer science but applied mathematics, the physical sciences, mechanical engineering, and the earth/space sciences. Once the project is completed, it will showcase both the qualitative and quantitative aspects of vector graphics and image modeling.

By employing an integrated approach of combining diffusion curve and gradient mesh, Qin will drastically expand the applied scope of vector graphics to visual information modeling, analysis, and processing, where numerical measurements are prevalent.

About the Researcher
Hong Qin is a professor in Stony Brook University’s Department of Computer Science and a member of the Center for Visual Computing. Qin received his PhD in computer science from the University of Toronto and was awarded the Open Doctoral Fellowship. During the past 20 years, Qin has secured more than 15 NSF awards, including the prestigious NSF CAREER Award in 1997 and two Information Technology Research (ITR) grants in 2001 and in 2003, respectively. In 2001, Qin also received the Sloan Fellowship.

-Samantha Mercado