Abstract: We present an automatic, real-time video and image abstraction framework that abstracts imagery by modifying the contrast of visually important features, namely luminance and color opponency. We reduce contrast in low-contrast regions using an approximation to anisotropic diffusion, and artificially increase contrast in higher contrast regions with difference-of-Gaussian edges. The abstraction step is extensible and allows for artistic or data-driven control. Abstracted images can optionally be stylized using soft color quantization to create cartoon-like effects with good temporal coherence. Our framework design is highly parallel, allowing for a GPU-based, real-time implementation. We evaluate the effectiveness of our abstraction framework with a user-study and find that participants are faster at naming abstracted faces of known persons compared to photographs. Participants are also better at remembering abstracted images of arbitrary scenes in a memory task.

Download Paper. Project Page. Watch Video. Bibtex.

Abstract: Visually important image features often disappear when color images are converted to grayscale. The algorithm introduced here reduces such losses by attempting to preserve the salient features of the color image. The Color2Gray algorithm is a 3-step process: 1) convert RGB inputs to a perceptually uniform CIE L*a*b* color space, 2) use chrominance and luminance differences to create grayscale target differences between nearby image pixels, and 3) solve an optimization problem designed to selectively modulate the grayscale representation as a function of the chroma variation of the source image. The Color2Gray results offer viewers salient information missing from previous grayscale image creation methods.

Download Paper. Project Page. Source Code and Implementation Notes. Bibtex.

Abstract: We present an interactive system which allows users to create abstract paintings in the style of Jackson Pollock using three dimensional viscous fluid jets. Pollock's paintings were created by using streams of household paint to make guided, semi-random patterns on his canvas. Our fluid jet model consists of two coupled simulations: a Navier-Stokes solver for an axis-symmetric fluid column and a linked-mass system for tracking the three dimensional motion of the jet's axis line. The paint trails left by the jets are represented using implicit surfaces. Our system also includes an algorithm for generating the splatter patterns created by the impacts of a highspeed fluid drops. We allow users to analyze the fractal properties of the images they create, comparing them to those known to exist in Pollock's own paintings.

Download Paper. Project Page. Download Video. Bibtex.

Abstract: We present techniques for generating and manipulating vector fields for use in the creation of painterly images and animations. Our aim is to enable casual users to create results evocative of expressionistic art. Rather than defining stroke alignment fields globally, we divide input images into regions using a colorspace clustering algorithm. Users interactively assign characteristic brush stroke alignment fields and stroke rendering parameters to each region. By combining vortex dynamics and semi-Lagrangian fluid simulation we are able to create stable, easily controlled vector fields. In addition to fluid simulations, users can align strokes in a given region using more conventional field models such as smoothed gradient fields and optical flow, or hybrid fields that combine the desirable features of fluid simulations and smoothed gradient information.

Download Paper. Bibtex.