The analysis and efficient implementation of direct binary search and related halftoning and image enhancement applications
Abstract
As the desktop printing market becomes more competitive, the demand for halftoning techniques that generate very high quality halftones is likely to increase. To achieve the highest level of print quality, a model based halftoning algorithm such as DBS must be used. This algorithm is an iterative method designed to minimize a metric of error between the grayscale original and halftone image. This metric incorporates a model for the human visual system (HVS) and the printer used to render the image. Unfortunately, this type of algorithm is required to employ an ad-hoc optimization strategy to solve this inherently nonlinear problem. As a result, these algorithms are extremely difficult to analyze. In addition, they are more computationally intensive than conventional halftoning techniques, and therefore less practical. First, DBS is analyzed to show why the algorithm produces halftones which exhibit high quality textures. This analysis is referred to as the Dual interpretation of DBS. Here, we have been able to establish that all halftones generated by DBS are guaranteed to satisfy a set of perceptually based performance criteria. Then, this analysis is applied to the design of dispersed dot screen functions. We compare and contrast DBS and Void-and-Cluster, two popular approaches for designing dither arrays. Next, we focus our attention on reducing the computational complexity of DBS. This effort includes a technique for implementing DBS on platforms with limited computing resources. Lastly, a perceptually based technique for sharpening images is provided. This approach is a direct outgrowth of the use of visual models in our prior work.
Degree
Ph.D.
Advisors
Allebach, Purdue University.
Subject Area
Electrical engineering
Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server.