Digital Image Processing and Machine Learning Research: Digital Color Halftoning, Printed Image Artifact Detection and Quality Assessment, and Image Denoising

In the thesis, we study several problems related to digital image processing and machine learning. Digital halftoning is an essential preprocessing step in printing. It is used to represent a continuous-tone image as an image that contains a limited number of colorants and with minimal loss of image quality. The primary goal of digital color halftones is to reproduce the closest possible color representation to the original image. To begin with, we describe a project in which three screens for Cyan, Magenta, and Yellow colorants were designed jointly using the Direct Binary Search algorithm (DBS). The screen set generated by the algorithm can be used to halftone color images easily and quickly. The halftoning results demonstrate that by utilizing the screen sets, it is possible to obtain high-quality color halftone images while significantly reducing computational complexity. The demands for printing speed and quality are increasing as imaging technology advances. Our next research focuses on defect detection and quality assessment of printed images, which are critical for designing and improving high-quality printing systems. We measure and analyze macro-uniformity, banding, and color plane misregistration, which are thought to have the greatest influence on printed image quality. For these three defects, we designed different pipelines for them and developed a series of digital image processing and computer vision algorithms for the purpose of quantifying and evaluating these printed image defects. Additionally, we conduct a human psychophysical experiment to collect perceptual assessments and use machine learning approaches to predict image quality scores based on human vision. Due to the limitations of various recording devices, images are sensitive to random noise during acquisition. Noise is a signal distortion that impedes image observation and information extraction. Thus, as a fundamental topic of image analysis and processing, image noise suppression aids our understanding of image statistics and processing. We study modern deep convolutional neural networks for image denoising and focus on blind, bias-removed, mix loss optimized, and perceptually oriented image denoising tasks. We propose a network designed for AWGN image denoising. Our network removes the bias at each layer to achieve the benefits of scaling invariant network; additionally, it implements a mix loss function to boost performance. We train and evaluate our denoising results using PSNR, SSIM, and LPIPS, and demonstrate that our results achieve impressive performance evaluated with both objective and subjective IQA metrics