Application of Zero-Watermarking Scheme Based on Swin Transformer for Securing the Metaverse Healthcare Data

Han, Baoru; Wang, Han; Qiao, Dawei; Xu, Jia; Yan, Tianyu

doi:10.1109/jbhi.2023.3257340

Cited by 10 publications

(3 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In the Swin Transformer network architecture, the Swin Transformer Block employs Windows Multi-Head Self-Attention (W-MSA) ( Li et al., 2021 ) and Shifted Window Multi-Head Self-Attention (SW-MSA) ( Han et al., 2023 ). The purpose of W-MSA is to reduce computational complexity.…”

Section: Principles and Methodsmentioning

confidence: 99%

Denoising Diffusion Probabilistic Models and Transfer Learning for citrus disease diagnosis

Li,

Guo,

Qiu

et al. 2023

Front. Plant Sci.

View full text Add to dashboard Cite

ProblemsPlant Disease diagnosis based on deep learning mechanisms has been extensively studied and applied. However, the complex and dynamic agricultural growth environment results in significant variations in the distribution of state samples, and the lack of sufficient real disease databases weakens the information carried by the samples, posing challenges for accurately training models.AimThis paper aims to test the feasibility and effectiveness of Denoising Diffusion Probabilistic Models (DDPM), Swin Transformer model, and Transfer Learning in diagnosing citrus diseases with a small sample.MethodsTwo training methods are proposed: The Method 1 employs the DDPM to generate synthetic images for data augmentation. The Swin Transformer model is then used for pre-training on the synthetic dataset produced by DDPM, followed by fine-tuning on the original citrus leaf images for disease classification through transfer learning. The Method 2 utilizes the pre-trained Swin Transformer model on the ImageNet dataset and fine-tunes it on the augmented dataset composed of the original and DDPM synthetic images.Results and conclusionThe test results indicate that Method 1 achieved a validation accuracy of 96.3%, while Method 2 achieved a validation accuracy of 99.8%. Both methods effectively addressed the issue of model overfitting when dealing with a small dataset. Additionally, when compared with VGG16, EfficientNet, ShuffleNet, MobileNetV2, and DenseNet121 in citrus disease classification, the experimental results demonstrate the superiority of the proposed methods over existing approaches to a certain extent.

show abstract

Section: Principles and Methodsmentioning

confidence: 99%

Denoising Diffusion Probabilistic Models and Transfer Learning for citrus disease diagnosis

Li,

Guo,

Qiu

et al. 2023

Front. Plant Sci.

View full text Add to dashboard Cite

show abstract

“…To optimize the feature space, they focused on eliminating redundancy by learning inter and intra-feature weights and incorporated a noise layer during feature training to increase robustness. Han et al [52] enhanced this methodology by introducing a chaotic encryption algorithm to encrypt the watermark before the XOR operation, enhancing security. They also adopted the Swin Transformer [53] to generate features for master share creation, achieving a feature space that is invariant to geometric distortions and enhances the robustness of the watermarking process.…”

Section: Methods Using Deep Network For Feature Transformationmentioning

confidence: 99%

A Brief, In-Depth Survey of Deep Learning-Based Image Watermarking

Zhong,

Das,

Alrasheedi

et al. 2023

Applied Sciences

View full text Add to dashboard Cite

This paper presents a comprehensive survey of deep learning-based image watermarking; this technique entails the invisible embedding and extraction of watermarks within a cover image, aiming for a seamless combination of robustness and adaptability. We navigate the complex landscape of this interdisciplinary domain, linking historical foundations, current innovations, and prospective developments. Unlike existing literature, our study concentrates exclusively on image watermarking with deep learning, delivering an in-depth, yet brief analysis enriched by three fundamental contributions. First, we introduce a refined categorization, segmenting the field into embedder–extractor, deep networks for feature transformation, and hybrid methods. This taxonomy, inspired by the varied roles of deep learning across studies, is designed to infuse clarity, offering readers technical insights and directional guidance. Second, our exploration dives into representative methodologies, encapsulating the diverse research directions and inherent challenges within each category to provide a consolidated perspective. Lastly, we venture beyond established boundaries, outlining emerging frontiers and providing detailed insights into prospective research avenues.

show abstract

“…In the same year, Zermi et al 29 proposed a blind watermarking method for medical image protection, which enhances security and ensures data integrity by adding electronic medical record hashes. And in 2023, Han et al 30 proposed a robust zero watermarking scheme based on the Swin transformer to improve the security of medical images in metadata healthcare systems. At the same time, privacy protection research is also progressing and developing [31][32][33] .…”

Section: Introductionmentioning

confidence: 99%

Study of research hotspots and development trends of image security under the background of deep learning

Guo,

Liu

2023

Preprint

View full text Add to dashboard Cite

As an important carrier of information transmission, images are widely used in human-computer interaction, digital twins, cloud computing, and even people’s daily life. Image security is directly related to personal privacy, business secrets and national security. It is of great significance to strengthen the research of image security. Deep learning is a machine learning method based on neural networks, which enables the processing and analysis of complex data by simulating the way human brain neurons work and provides strong support for research in the field of image security. With the rapid development of deep learning technology, its application in the field of image processing and security is increasingly extensive. To gain a deeper understanding of the current situation and cutting-edge trends of deep learning-based image security research. This paper uses CiteSpace, a bibliometric tool, to analyze the 2009-2022 research results on "image security" and "deep learning" in the core database of Web of Science. Accurately grasp the stage of development and characteristics of image security research in the world through data analysis. It is found that the research hotspots mainly cover image encryption, image steganography, image watermarking, and other technical applications. In the past decade, great progress has been made in the field of image security. The research on techniques such as synthetic image identification, image tampering detection, and adversarial samples is rapidly emerging. Future research on image security deep learning-based will focus on the directions of anti-attacks, robustness enhancement, privacy protection, and data sharing. We hope this study will provide new research ideas and thinking perspectives to draw on for the future direction of image security.

show abstract

Application of Zero-Watermarking Scheme Based on Swin Transformer for Securing the Metaverse Healthcare Data

Cited by 10 publications

References 0 publications

Denoising Diffusion Probabilistic Models and Transfer Learning for citrus disease diagnosis

Denoising Diffusion Probabilistic Models and Transfer Learning for citrus disease diagnosis

A Brief, In-Depth Survey of Deep Learning-Based Image Watermarking

Study of research hotspots and development trends of image security under the background of deep learning

Contact Info

Product

Resources

About