C.-C. Chang, C.-S. Chan, and Y.-P. Pan (Taiwan)
halftone image, error diffusion, LMSadaptive filtering
In this paper, we shall propose a method to inverse halftone images. Due to the low memory cost for filter masking, we construct our inverse halftoning technique by combining the least-mean-square (LMS) inverse halftoning technique and the genetic algorithm to produce filter masks. Our method includes the following two steps. First, we use the LMS adaptive filtering to produce the coarse filter masks for each image and eliminate the least significant weight to form the global mask shape. Second, we use the genetic algorithm to find the good weights on the mask shape acquired from the first step. After getting the good weights, we can use them to reconstruct gray level images from halftone images. The application results of our method are also shown in this paper in the situation of low memory cost.
Important Links:
Go Back
