Towards High-Resolution Copy-Evident Ceramic Tiles: A Deep Learning Framework for Halftoning and Watermarking

Ceramic art is essential in interior design and decoration, and making exquisite ceramic tiles imposes strict requirements for inkjet printing technology. High-resolution ceramic tiles are often produced through inkjet printing, in which the input images are converted into a halftone format. However, traditional binary halftoning techniques cannot produce high-resolution images for the ensuing printing process. Given that the processes of inkjet printing and high-temperature firing of ceramic tiles are a highly complex nonlinear system, and existing halftoning methods pose intractable problems, including inconsistent textures and color deviations. Based on a modified U-Net model and a modified error diffusion algorithm, we propose a multilevel halftoning method, which is capable of converting color-separation images of ceramic tiles into high-resolution halftone images. To deter copyright infringement, we further apply an ad hoc invisible watermarking method for halftone images. In this paper, we propose a four-stage framework: (1) A self-built dataset is used to solve non-convergence and overfitting problems caused by the unbalanced samples and non-uniqueness of halftone images. (2) A modified U-Net model is trained on the self-built dataset and applied to the ceramic-tile images. (3) An improved error diffusion algorithm is used to calibrate and convert the predicted continuous-tone transition images into multilevel halftone images for inkjet printing. (4) A invisible and robust watermark is embedding algorithm towards halftone images is proposed for copyright protection. Experimental results show that our methodology is effective for performing the color-to-halftone transformation and identifying the copyright.

Layer Decomposition Learning Based on Gaussian Convolution Model and Residual Deblurring for Inverse Halftoning

Layer decomposition to separate an input image into base and detail layers has been steadily used for image restoration. Existing residual networks based on an additive model require residual layers with a small output range for fast convergence and visual quality improvement. However, in inverse halftoning, homogenous dot patterns hinder a small output range from the residual layers. Therefore, a new layer decomposition network based on the Gaussian convolution model (GCM) and a structure-aware deblurring strategy is presented to achieve residual learning for both the base and detail layers. For the base layer, a new GCM-based residual subnetwork is presented. The GCM utilizes a statistical distribution, in which the image difference between a blurred continuous-tone image and a blurred halftoned image with a Gaussian filter can result in a narrow output range. Subsequently, the GCM-based residual subnetwork uses a Gaussian-filtered halftoned image as the input, and outputs the image difference as a residual, thereby generating the base layer, i.e., the Gaussian-blurred continuous-tone image. For the detail layer, a new structure-aware residual deblurring subnetwork (SARDS) is presented. To remove the Gaussian blurring of the base layer, the SARDS uses the predicted base layer as the input, and outputs the deblurred version. To more effectively restore image structures such as lines and text, a new image structure map predictor is incorporated into the deblurring network to induce structure-adaptive learning. This paper provides a method to realize the residual learning of both the base and detail layers based on the GCM and SARDS. In addition, it is verified that the proposed method surpasses state-of-the-art methods based on U-Net, direct deblurring networks, and progressively residual networks.

Inverse Halftoning Based on Weighted Nuclear Norm Minimization

The inverse halftoning technology refers to restore a continuous-tone image from a halftone image with only bi-level pixes. However, recovering the continuous images from their halftoned ones is normally ill-posed, which making the inverse halftoning algorithm very challenging. In this paper, we propose an optimization model with two alternate projections (TAP) for image inverse halftoning under the weighted nuclear norm minimization (WNNM) framework. The main contributions are two-folds. First, the WNNM nonlocal regularization term is established, which offers a powerful mechanism of nonlocal self-similarity to ensure a more reliable estimation. Second, the alternate minimization projections are formulated for solving the image inverse halftoning, which reconstructs the continuous-tone image without destroying the image details and structures. The experiment results shown that the proposed method outperformed the state of the arts in terms of both objective measurements and subjective visual performance.

Improved Content-Color-Dependent Screening (CCDS): Adaptive Bilateral Filtering and Color-Aware Sobel Edge Detector

In previous work [1] , content-color-dependent screening (CCDS) determines the best screen assignments for either regular or irregular haltones to each image segment, which minimizes the perceived error compared to the continuous-tone digital image. The model first detects smooth areas of the image and applies a spatiochromatic HVS-based model for the superposition of the four halftones to find the best screen assignment for these smooth areas. The segmentation is not limited to separating foreground and background. Any significant color regions need to be segmented. Hence, the segmentation method becomes crucial. In this paper, we propose a general segmentation method with a few improvements: The number of K-means clusters is determined by the elbow method to avoid assigning the number of clusters manually for each image. The noise removing bilateral filter is adaptive to each image, so the parameters do not need to be tested and adjusted based on the visual output results. Also, some color regions can be clearly separated out from other color regions by applying a color-aware Sobel edge detector.

Structure-Aware Halftoning Using the Iterative Method Controlling the Dot Placement

Many image reproduction devices, such as printers, are limited to only a few numbers of printing inks. Halftoning, which is the process to convert a continuous-tone image into a binary one, is, therefore, an essential part of printing. An iterative halftoning method, called Iterative Halftoning Method Controlling the Dot Placement (IMCDP), which has already been studied by research scholars, generally results in halftones of good quality. In this paper, we propose a structure-based alternative to this algorithm that improves the halftone image quality in terms of sharpness, structural similarity, and tone preservation. By employing appropriate symmetrical and non-symmetrical Gaussian filters inside the proposed halftoning method, it is possible to adaptively change the degree of sharpening in different parts of the continuous-tone image. This is done by identifying a dominant line in the neighborhood of each pixel in the original image, utilizing the Hough Transform, and aligning the dots along the dominant line. The objective and subjective quality assessments verify that the proposed structure-based method not only results in sharper halftones, giving more three-dimensional impression, but also improves the structural similarity and tone preservation. The adaptive nature of the proposed halftoning method makes it an appropriate algorithm to be further developed to a 3D halftoning method, which could be adapted to different parts of a 3D object by exploiting both the structure of the images being mapped and the 3D geometrical structure of the underlying printed surface.

Texture-aware Error Diffusion Algorithm for Multi-level Digital Halftoning

Digital halftoning is a technique for converting a continuous-tone image into a quantized image to reproduce it on a digital printing device. Error diffusion (ED) is an algorithm that has proven to be effective for the halftoning process, and it has been widely applied to digital printing tasks. However, in images reproduced using conventional ED algorithms based on the signal processing theory, the texture of objects is often lost. In this study, we propose a texture-aware ED algorithm for multi-level digital halftoning. First, we generate multiple mapped images with different brightness levels through nonlinear transformation. For each mapped image, we adopt a texture-aware binary error diffusion method to obtain multiple halftone images. Finally, we generate a multi-level halftone image from the multiple halftone images. We test the algorithm on an actual printer, compare the results with those of the current raster image processor software and classical ED algorithms, and observe that our algorithm outputs better results.

The Polychromatic Woodburytype—Colour Tracking in Translucent, Patterned Gelatin/Pigment Films

The Woodburytype is a 19th century photomechanical technique capable of producing high-quality continuous-tone prints. It uses pigment dispersed in gelatin to produce a 2.5D print, in which the effect of varying tone is produced by a variation in the print height. We propose a method of constructing full colour prints in this manner, using a CMY colour model. This involves the layering of multiple translucent pigmented gelatin films and tracking how the perceived colour of these stacks changes with varying height. A set of CMY inks is constructed, taking into account the optical properties of both the pigment and gelatin, and a method of translating images into these prints is detailed.

Inverse Halftoning Methods Based on Deep Learning and Their Evaluation Metrics: A Review

Inverse halftoning is an ill-posed problem that refers to the problem of restoring continuous-tone images from their halftone versions. Although much progress has been achieved over the last decades, the restored images still suffer from detail loss and visual artifacts. Recent studies show that inverse halftoning methods based on deep learning are superior to other traditional methods, and thus this paper aimed to systematically review the inverse halftone methods based on deep learning, so as to provide a reference for the development of inverse halftoning. In this paper, we firstly proposed a classification method for inverse halftoning methods on the basis of the source of halftone images. Then, two types of inverse halftoning methods for digital halftone images and scanned halftone images were investigated in terms of network architecture, loss functions, and training strategies. Furthermore, we studied existing image quality evaluation including subjective and objective evaluation by experiments. The evaluation results demonstrated that methods based on multiple subnetworks and methods based on multi-stage strategies are superior to other methods. In addition, the perceptual loss and the gradient loss are helpful for improving the quality of restored images. Finally, we gave the future research directions by analyzing the shortcomings of existing inverse halftoning methods.

Effects of Dot Configuration on Color Reproduction and Anti-counterfeiting Features

This study represents an attempt to solve the problem of color reproduction and identification for the prevention of stamp forgery. Generally, printed images are converted to halftone dot patterns by using a raster image processor. The amplitude-modulated dots of each separated color plane have a consistent shape such as conventional round, square, ellipse, or diamond shape; error diffusion occurs in frequency-modulated dots. To achieve anti-counterfeiting properties for stamp reproduction, two methods are proposed to obtain difficult-to-replicate dot structures and to provide corresponding color management methods. The first method involves arranging different dot shapes in different areas of an image. Color consistency is achieved using a virtual gray balance method. However, color differences are visible when two dot types are assigned to adjacent areas with similar colors. The second method is a two-stage screening method. The first screening stage defines different micro-regions in the image, which are then combined with the continuous-tone image again in the second screening step to assign different dot patterns to different micro-regions. This approach not only provides anti-counterfeiting ability but also ensures color consistency and allows controlling color quality using one ICC profile.

Reversible Data Hiding in Encrypted Color Halftone Images with High Capacity

In recent years, reversible data hiding (RDH) has become a research hotspot in the field of multimedia security that has aroused more and more researchers’ attention. Most of the existing RDH algorithms are aiming at continuous-tone images. For RDH in encrypted halftone images (RDH-EH), the original cover image cannot be recovered losslessly after the watermark is extracted. For some application scenarios such as medical or military images sharing, reversibility is critical. In this paper, a reversible data hiding scheme in encrypted color halftone images (RDH-ECH) is proposed. In the watermark embedding procedure, the cover image is copied into two identical images to increase redundancy. We use wet paper code to embed the watermark into the image blocks. Thus, the receiver only needs to process the image blocks by the check matrices in order to extract the watermarks. To increase embedding capacity, we embed three layers in the embedding procedure and combine the resulting images into one image for convenience of transmission. From the experimental results, it can be concluded that the original image can be restored entirely after the watermarks are extracted. Besides, for marked color halftone images, our algorithm can implement high embedding capacity and moderate visual quality.