Algorithms for Adaptive Palette Construction using the Graphics Processing Unit
Date of Award
Doctor of Philosophy (PhD)
Graduate School of Computer and Information Sciences
Michael J. Laszlo
Computer image processing can be costly in terms of computing resources. Two image characteristics that can contribute to this cost, resolution and color depth, are essential when considering such tasks as changing an image's resolution or compressing the image for optimal storage. Even with the introduction of a number of algorithms to address the high cost of image processing, image processing continues to incur a penalty in terms of efficiency. Adaptive color palettes represent one approach that addresses the problem of decreased efficiency in image processing by quantizing an image from one color space to another. Additionally, recent advances have yielded a way to take advantage of the processing capabilities of a computer's graphics processing unit (GPO). These GPU advances provide yet another way to address the cost of image processing by making use of general purpose processing on the GPU. By using quantizing algorithms in conjunction with general-purpose GPU algorithms, it is possible to identify algorithmic approaches that would be applicable to GPGPU processing with the purpose of decreasing the cost of processing color image palettes.
Victor Johnson. 2008. Algorithms for Adaptive Palette Construction using the Graphics Processing Unit. Doctoral dissertation. Nova Southeastern University. Retrieved from NSUWorks, Graduate School of Computer and Information Sciences. (616)