Multi-scale texture loss for CT denoising with GANs

Abstract: Restoring damaged character photographs (DCPs) is essential for preserving cultural and personal memories, yet traditional methods are labor-intensive and often require significant manual intervention. The degradation of photographs due to factors such as creases, spots, water damage, and exposure to light poses a considerable challenge to conventional restoration techniques. Manual restoration, typically performed using software like Photoshop, is not only time-consuming but also demands a high level of expertise and meticulous effort, making it impractical for large-scale applications. In this study, we propose an innovative solution to automate the restoration process by employing a practical generative adversarial network (GAN) architecture. Our approach addresses the limitations of traditional methods and aims to significantly reduce the time and effort required for photo restoration while improving the quality of the restored images. The proposed method comprises a novel three-step pipeline designed to tackle the complexities of DCPs effectively. First, we initiate the process by collecting a diverse dataset, including clear character photographs (CCPs), real DCPs, and dirty masks that illustrate common damage patterns. This comprehensive dataset forms the foundation for training our models. Second, we utilize a Residual U-Net GAN (RUGAN) to learn the spoilage patterns of DCPs and generate realistic fake DCPs from the CCPs and dirty masks. RUGAN leverages the structural and textural information from the clear images and dirty masks to produce fake damaged images that closely mimic real-world damage. Finally, we train a restoration model known as the Residual U-Net conditional GAN (RUCGAN) using the paired fake DCPs and CCPs.

Single Image Super-Resolution (SISR) stands at the forefront of image processing, offering a transformative solution to elevate low-resolution images to a higher fidelity.Leveraging deep neural networks, particularly Generative Adversarial Networks (GANs), this study aims to enrich the quality of images, especially in domains like medical imaging, public surveillance, and historical image restoration. Traditional methods often fail to preserve crucial details during up scaling, prompting the adoption of deep learning techniques which have demonstrated remarkable breakthroughs in accuracy and efficiency. By harnessing the adversarial training process inherent in GANs, we seek to propel the boundaries of image enhancement, striving to faithfully reconstruct high-resolution renditions while retaining essential nuances from their low-resolution counterparts. Through comprehensive performance evaluations utilizing publicly available datasets, we conduct rigorous quantitative and qualitative analyses to gauge the efficacy of various GAN architectures in SISR. This meticulous examination allows us to assess not only the fidelity of the reconstructed images but also the computational efficiency and scalability of each approach, essential considerations for real-world applications. However, alongside the promises of GAN-based SISR come inherent challenges that warrant attention. Issues such as training instability, mode collapse, and imageartifacts present hurdles that must be addressed to fully harness the potential of GANs in image enhancement.Additionally, the integration of domain-specific expertise, particularly in fields like medical imaging, can further enhance the utility and efficacy of GAN-based super-resolution techniques.In conclusion, while the journey towards achieving high-quality, high-resolution images may be fraught with challenges, the integration of GANs into SISR represents a significant leap forward, promising to revolutionize various industries and deepen our understanding of visual data.

Image dehazing has always been a challenging topic in image processing. The development of deep learning methods, especially the generative adversarial networks (GAN), provides a new way for image dehazing. In recent years, many deep learning methods based on GAN have been applied to image dehazing. However, GAN has two problems in image dehazing. Firstly, For haze image, haze not only reduces the quality of the image but also blurs the details of the image. For GAN network, it is difficult for the generator to restore the details of the whole image while removing the haze. Secondly, GAN model is defined as a minimax problem, which weakens the loss function. It is difficult to distinguish whether GAN is making progress in the training process. Therefore, we propose a guided generative adversarial dehazing network (GGADN). Different from other generation adversarial networks, GGADN adds a guided module on the generator. The guided module verifies the network of each layer of the generator. At the same time, the details of the map generated by each layer are strengthened. Network training is based on the pre-trained VGG feature model and L1-regularized gradient prior which is developed by new loss function parameters. From the dehazing results of synthetic images and real images, the proposed method is better than the state-of-the-art dehazing methods.

Rice is a staple food for more than half of the world’s population. Furthermore, rice is the main export crop of Thailand which produces 21% world’s market share. However, weedy rice is a major counterproductive plant that reduces rice productivity by more than 80% in Thailand. Previous research attempted to develop image classification models to recognize types of rice using images captured in closed environments, which is not practical for farmers with typical mobile phone cameras. This research develops a specific Generative Adversarial Network (GAN) architecture to translate an input image from a typical mobile phone cameras into the closed environment setting. Our GAN architecture can translate mobile phone images and achieves 90.06% weedy rice recognition accuracy, as compared to 58.10% without the translation.

Generative adversarial networks (GANs) have been widely investigated for many potential applications in medical imaging. DatasetGAN is a recently proposed framework based on modern GANs that can synthesize high-quality segmented images while requiring only a small set of annotated training images. The synthesized annotated images could be potentially employed for many medical imaging applications, where images with segmentation information are required. However, to the best of our knowledge, there are no published studies focusing on its applications to medical imaging. In this work, preliminary studies were conducted to investigate the utility of DatasetGAN in medical imaging. Three improvements were proposed to the original DatasetGAN framework, considering the unique characteristics of medical images. The synthesized segmented images by DatasetGAN were visually evaluated. The trained DatasetGAN was further analyzed by evaluating the performance of a pre-defined image segmentation technique, which was trained by the use of the synthesized datasets. The effectiveness, concerns, and potential usage of DatasetGAN were discussed.

Generative AI is an expanding domain that employs machine learning models to generate novel data that closely mimic pre existing data. ChatGPT and DALL-E can be customized for specific applications and are expected to transform healthcare, education, and communication. Generative Adversarial Networks (GANs) that can generate synthetic medical images closely mimicking actual patient data may substantially enhance machine learning model training datasets. They also translate medical images from one modality to another, improve medical imaging resolution, reduce radiation exposure, and boost image quality and detail. Despite their challenges, GANs have great potential in the field of medical imaging. The key obstacles are the need for Graphic Processing Units (GPUs) and computing resources to train GANs and the lack of established standards for generating synthetic images. Incorrectly labeled data for training other machine learning models can reduce performance, making ground-truth data labeling for healthcare AI more difficult. Generative AI is revolutionizing healthcare imaging, simplifying diagnosis, and propelling healthcare research and practice to new frontiers. Ensuring the reliability and safety of generated images in medical applications requires addressing ethical considerations and validating data.

Rapid imaging of subsurface electrical structures is highly challenging, especially for complex geological formations. Conventional inversion algorithms require repeated solutions of large-scale forward problems, which constitute the main computational expense. To address this limitation, we have developed an underground resistivity imaging method based on the Pix2Pix Generative Adversarial Network (GAN) architecture. Our approach integrates impedance phase information with conventional apparent resistivity observations, significantly improving imaging accuracy. For training data generation, we employ Gaussian random fields to synthesize resistivity models. This practice not only enhances the geological representativeness of the data but also introduces meaningful variability that benefits the generalization capability of the GAN. By systematically comparing the prediction accuracy under different loss functions, we determined the optimal form of the loss function.Detailed qualitative and quantitative evaluations demonstrate that our multi-parameter joint inversion strategy outperforms methods relying on only a single parameter, such as apparent resistivity or impedance phase alone. To improve the method’s robustness in practical applications, we incorporate the objective function from conventional inversion into the GAN’s loss function to handle noisy data. This geophysically constrained loss function greatly enhances the model’s noise resistance. In synthetic data experiments, compared with the Nonlinear Conjugate Gradient (NLCG) inversion method, our approach not only achieves faster prediction but also exhibits superior capability in resolving high-resistivity bodies beneath low-resistivity layers. Validation using real-world data further confirms the practical applicability and generalization potential of the proposed method.

This paper provides an in-depth review of recent advancements in image super-resolution (SR) techniques using deep learning, with a specific focus on their applications in medical imaging. In the medical field, high-resolution images are crucial for accurate diagnosis and treatment planning. However, due to hardware limitations, it is often challenging to obtain high-resolution (HR) medical images. Traditional image enhancement techniques are not always sufficient to recover fine details. Deep learning-based methods, particularly those leveraging convolutional neural networks (CNNs), generative adversarial networks (GANs), and transformer-based models, have shown remarkable success in overcoming these challenges. The review explores the evolution of these techniques, their advantages, and the challenges involved in medical image super-resolution. The paper also highlights the future directions of research to improve SR for medical imaging, addressing concerns related to image quality, data availability, and computational efficiency.

Generative adversarial networks(GANs) have been successfully applied for generating high quality natural images and have been extended to the generation of RGB videos and 3D volume data. In this paper we consider the task of generating RGB-D videos, which is less extensively studied and still challenging. We explore deep GAN architectures suitable for the task, and develop 4 GAN architectures based on existing video-based GANs. With a facial expression database, we experimentally find that an extended version of the motion and content decomposed GANs, known as MoCoGAN, provides the highest quality RGB-D videos. We discuss several applications of our GAN to content creation and data augmentation, and also discuss its potential applications in behavioral experiments.

Embodying the design from scratches has been a common problem for all designers. In some situation, clothes designers have trouble creating works that match their idea the most. Creating a fine work from a simple sketch is a tedious process of endless try-and-error. Designers spend days, weeks, or even months drawing drafts and discard them, because they need to draw the picture that best embodies the idea in their head. This study proposes a deep learning method that can generate clothes images from human sketch input, explores the effect of different image to sketch translators, different Generative Adversarial Network (GAN) architecture, loss functions and regularization methods. Many previous works have demonstrated strong results on conditioned image generation, and this study shows that similar settings can be transferred to help designers to embody their designs. The result shows that the model can generate clothes images with high fidelity while faithful to the input sketch.

Purpose: The purpose of this article is to offer an overview of the recent applications of cloud computing in medical imaging and their advances in the field. It reviews existing scientific and academic literature on cloud computing-based platforms and solutions for medical image analysis and storage of medical imaging data. Materials and methods: This article uses available scientific literature on the applications of cloud computing in medical imaging from PubMed, Google Scholar and ScienceDirect. Results: The review shows that interest and research in cloud computing applications in medical imaging has increased in recent years. This has led to new and more effective platforms and solutions for medical image analysis and storage of medical imaging data. Innovative applications of cloud computing in medical imaging try to address ethical and security concerns using authentication, encryption, anonymization and access controls. Conclusions: Cloud computing offers promising applications in medical imaging. Notwithstanding, further research is necessary to demonstrate their effectiveness, safety and security in a medical setting.

Text, as a vital tool for communication, is playing an imperative role in modern society. Precise high-level text translation systems are essential requirements in a wide range of real-world applications, such as robot navigation, industrial automation, image search, and instant translation. Regardless of improved research, a series of grand challenges may still become upon when translating text automatically in the real-world from open scene images. The difficulties mainly stem from multiplicity and inconsistency of text in open scenes, complication and obstruction of backgrounds, and deficient imaging conditions in uncontrolled circumstances for open scene images. The existing deep learning-based text translation systems do not eliminate the text for translation, and these applications just replace text on the reconstructed scene. To address the abovementioned shortcomings, this study proposed a novel approach for open scene text translation. Our system consists of five modules including scene text detection, text recognition, text elimination, text translation, and text insertion along with scene reconstruction. The novelty presented by our model lies in the idea of first eliminating the text from the open scene for accurate translation and then reconstructs the translated text on the image for its proper alignment. We specifically modified the existing generative adversarial network (GAN) architecture for improved performance of text elimination by introducing a novel strategy of text and scene concatenation to reduce the overall loss function. For this purpose, we created a synthetic dataset to train our GAN for text elimination module. Experiments on various standard text translation systems demonstrate that our integrated system is able to outperform state-of-the-art approaches in terms of result quality. We have achieved 90.87% of precision, 83.66% of recall, 87.116% of F1-score, and reduced both losses (l1 and l2) up to 50% which is remarkable upon state-of-the-art translation systems.

Shortly after deep learning algorithms were applied to Image Analysis, and more importantly to medical imaging, their applications increased significantly to become a trend. Likewise, deep learning applications (DL) on pulmonary medical images emerged to achieve remarkable advances leading to promising clinical trials. Yet, coronavirus can be the real trigger to open the route for fast integration of DL in hospitals and medical centers. This paper reviews the development of deep learning applications in medical image analysis targeting pulmonary imaging and giving insights of contributions to COVID-19. It covers more than 160 contributions and surveys in this field, all issued between February 2017 and May 2020 inclusively, highlighting various deep learning tasks such as classification, segmentation, and detection, as well as different pulmonary pathologies like airway diseases, lung cancer, COVID-19 and other infections. It summarizes and discusses the current state-of-the-art approaches in this research domain, highlighting the challenges, especially with COVID-19 pandemic current situation.

Generative AI has revolutionized computer vision by enabling machines to synthesize and enhance visual data. Advances in deep learning, such as Generative Adversarial Networks (GANs), Variational Autoencoders (VAEs), and diffusion models, have led to high-quality image generation with applications in medical imaging, data augmentation, and autonomous systems. Despite its potential, challenges like ethical concerns, dataset biases, and computational costs remain critical for future research and implementation. This paper explores the evolution of generative AI, its methodologies, applications, and the ethical and computational challenges it presents.

Introduction; The protection of patient privacy through medical image de-identification stands as a vital yet complicated healthcare challenge which demands both accurate diagnosis and privacy protection. Deep learning techniques now provide superior methods to enhance medical images which suffer from acquisition noise and low-resolution degradation. The research develops a Generative Adversarial Network (GAN) architecture to create a deep-learning solution which enhances medical images while removing identifying information.Objective; The proposed method uses adversarial learning to eliminate noise while restoring detailed high-resolution content from low-quality medical images.Method; The research team analysed medical images through data analysis of their dataset. The GAN model received training and validation through experiments that compared its performance against established demo denoising and super-resolution techniques to assess its overall performance. The fundamental technology shows promise for future medical applications because it can enhance image quality to diagnose and treat multiple diseases. Medical image analysis requires images with diverse detailed features for proper evaluation.Results; The proposed model demonstrates successful background noise reduction and improved image clarity with preserved diagnostic elements according to the obtained results. The proposed method achieved better results in PSNR and SSIM metrics than baseline models which proved its ability to restore vital diagnostic details. Conclusion; The research introduces an innovative GAN-based system which delivers superior medical image quality while maintaining patient information confidentiality during de-identification processes. The method shows promise to create efficient and economical diagnostic processes through its ability to analyze poor-quality medical images.