China
Africa
Explore chapters and articles related to this topic
Digital Halftoning
Published in Edward R. Dougherty, Digital Image Processing Methods, 2020
Paul G. Roetling, Robert P. Loce
Thus far, the halftone methods considered in detail were point processes. Here, we describe the neighborhood process of error diffusion, which uses the concept of fixing the total gray content of the image by calculating the brightness error incurred upon binarizing a pixel and incorporating this error in the processing of subsequent pixels. Due to resulting isolated white and isolated black pixels produced by the basic algorithm, the application of error diffusion has been primarily in display technologies (certain variations of the method do provide some clustering of like pixels, thereby rendering a printable image). The error diffusion method mitigates the trade-off of screen visibility versus gray-level contouring that is inherent in clustered-dot ordered-dither methods. Smoothly varying gray-scale images as well as sharp discontinuities are well rendered. A pleasant “blue noise” granular structure is observed with the exception of some undesirable worm-shaped artifacts. An example of error diffusion was shown in Fig. 8, where the pixel resolution is one quarter of that used in the clustered-dot examples so as to render the images printable and show the method at roughly the resolution of typical use. We discuss the basic algorithm introduced by Floyd and Steinberg [27] and some of its modifications, such as the use of other masks, perturbed thresholds, randomly distributed error, edge enhancement, and so on. To develop a deeper understanding of the fundamental technique, a one-dimensional frequency modulation model and a spectral analysis of the general technique are presented.
Color quantization
Published in Sharma Gaurav, Digital Color Imaging Handbook, 2017
Several investigations have shown that the error diffusion algorithm presents some major drawbacks. For example, according to Fletcher27 and Bouman,12 the colormap selection algorithm may be altered by the dithering process when input colors lie outside the convex hull of the representative colors. It is thus necessary to guarantee that all colors in the original image may be generated by taking a linear combination of colors in the colormap. Likewise, according to Knuth,41 another major drawback of the error diffusion algorithm is its inherent serial property; e.g., the value I′(x, y) depends in all pixels of input data I(x, y). Furthermore, it sometimes puts “ghosts” into the picture. Even if the ghosting problem can be ameliorated by choosing the coefficients Cij so that their sum is less than 1, the ghosts cannot be exorcised completely in this way. Finally, according to Bouman, 12when the error diffusion algorithm is used in conjunction with an unstructured color palette, it requires an exhaustive search of the closest representative (Section 9.7). Although some basic improvements to the exhaustive search method have been made (Section 9.7.1), this method remains computationally intensive.
Portable Network Graphics Approach to the Authentication of Halftone Images with Henon Map Encryption
Published in Smart Science, 2020
G. RajKumar, G. Udhaya Sankar, G. Ravi, C. Ganesa Moorthy, S. Sekar
The source image is converted into PNG format by adding alpha channel to the image. The entire process is divided into two sections. In the first section, the authentication and repairing data have been generated from the source image. In the second section, the stego image has been created followed by encryption. The image is first converted into binary format by using dot-diffused halftoning algorithm [13]. Dot diffusion is a halftoning technique which provides parallelism. In this, the whole image is divided into several nonoverlapped blocks and each block is processed individually. The processing is similar to error diffusion in which error is diffused to neighboring pixels. The halftoned image acts as the input for authentication signal generation. First, the halftoned image is divided into several blocks. For each block, authentication data have been generated. Hence, authentication is checked at the block level. If it fails, then the entire block is treated as tampered. The data for reconstruction are generated by using Shamir secret sharing algorithm [14]. Secret sharing has broad applications in the real world and can be used for situations in which access to important resources has to be protected. These data are then embedded into the alpha channel plane of the PNG image. This image is then encrypted by using Henon map to prevent unauthorized reading. The resulting image is called stego image. The block diagram of the proposed method is as shown in Figure 1.