Showing papers by "Charles R. Dyer published in 2007"

A text-to-picture synthesis system for augmenting communication

Abstract: We present a novel Text-to-Picture system that synthesizes a picture from general, unrestricted natural language text. The process is analogous to Text-to-Speech synthesis, but with pictorial output that conveys the gist of the text. Our system integrates multiple AI components, including natural language processing, computer vision, computer graphics, and machine learning. We present an integration framework that combines these components by first identifying infonnative and 'picturable' text units, then searching for the most likely image parts conditioned on the text, and finally optimizing the picture layout conditioned on both the text and image parts. The effectiveness of our system is assessed in two user studies using children's books and news articles. Experiments show that the synthesized pictures convey as much infonnation about children's stories as the original artists' illustrations, and much more information about news articles than their original photos alone. These results suggest that Text-to-Picture synthesis has great potential in augmenting human-computer and human-human communication modalities, with applications in education and health care, among others.

Patch-based Image Correlation with Rapid Filtering

Abstract: This paper describes a patch-based approach for rapid image correlation or template matching. By representing a template image with an ensemble of patches, the method is robust with respect to variations such as local appearance variation, partial occlusion, and scale changes. Rectangle filters are applied to each image patch for fast filtering based on the integral image representation. A new method is developed for feature dimension reduction by detecting the "salient" image structures given a single image. Experiments on a variety images show the success of the method in dealing with different variations in the test images. In terms of computation time, the approach is faster than traditional methods by up to two orders of magnitude and is at least three times faster than a fast implementation of normalized cross correlation.

Population Monte Carlo Path Tracing

Abstract: We present a novel global illumination algorithm which distributes more image samples on regions with perceptually high variance. Our algorithm iterates on a population of pixel positions used to estimate the intensity of eac h pixel in the image. A member kernel function, which automatically adapts to approximate the target ditribution b y using the information collected in previous iterations, is responsible for proposing a new sample position from the current one during the mutation process. The kernel function is designed to explore a proper area around the population sample to reduce the local variance. The resampling process eliminates samples located in the low-variance or well-explored regions and generates new samples to achieve ergocity. New samples are generated by considering two factors: the perceptual variance and the stratification of the sample distributions on the image plane. Our results show that the visual quality of the rendered image can be improved by exploring the correlated information among