Robust and Secure Watermarking of Medical Images Using Möbius Transforms

Al-Rammahi, Atheer; Sajedi, Hedieh

doi:10.1109/qicar61538.2024.10496638

Cited by 1 publication

(1 citation statement)

References 32 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…As healthcare systems become more digital and medical imaging tools like X-rays, MRIs, and CT scans become more common, it is more important than ever to make sure that these images are real. It is very important to find any possible changes or fakes in medical images because they have a direct effect on patient care, professional decision-making, and medical study [4].Traditional ways of verifying images, like watermarking and digital fingerprints, aren't very good at finding complex changes and fakes. Because of this, there is more and more interest in using cutting edge technologies, especially deep learning, to solve this problem.…”

Section: A Backgroundmentioning

confidence: 99%

Analytical Image Authentication in Healthcare: Deep Learning Based Mathematical Methods for Uncovering Large-Scale Forgery in Medical Images

Sharda Mahajan

2024

cana

View full text Add to dashboard Cite

Analytical Image Authentication in Healthcare uses a variety of deep learning methods to demonstrate a more advanced way of verifying images. It uses ResNets, Capsule Networks, Long Short-Term Memory (LSTM), Generative Adversarial Networks (GANs), and Convolutional Neural Networks (CNNs). Combining these state-of-the-art algorithms makes the system more accurate and reliable at finding large-scale forgery in medical pictures. In response to the growing danger of digitally changing healthcare images, this study creates a new, more thorough approach that goes beyond current methods. Forgery identification depends on CNNs, which make it possible to pull out complex picture data. Additionally, ResNets add more detail to models, which makes it easier to spot subtle trends that point to tampering. Using Capsule Networks gives the model a new point of view and lets it store structured connections within pictures, which improves its ability to identify things. According to research, LSTM networks help the system understand time better, which is important for finding small changes between medical scans. Additionally, using GANs adds a special competitive element that helps the model tell the difference between real and fake pictures through training against other models. After a lot of testing and improvement, the suggested way works better than others at finding fake activities in medical images. Utilizing cutting-edge deep learning methods and mathematical models together, this method guarantees the accuracy and trustworthiness of medical data, which builds trust in healthcare systems. To test the suggested way, a normal set of medical picture datasets with various types of fakes are used. These include copy-move, cutting, and editing. Experiments show that the multi-modal method is a good way to find and pinpoint faked areas because it is very accurate and not easily duplicated by different types of fraud.

show abstract

Section: A Backgroundmentioning

confidence: 99%

Analytical Image Authentication in Healthcare: Deep Learning Based Mathematical Methods for Uncovering Large-Scale Forgery in Medical Images

Sharda Mahajan

2024

cana

View full text Add to dashboard Cite

show abstract

Robust and Secure Watermarking of Medical Images Using Möbius Transforms

Cited by 1 publication

References 32 publications

Analytical Image Authentication in Healthcare: Deep Learning Based Mathematical Methods for Uncovering Large-Scale Forgery in Medical Images

Analytical Image Authentication in Healthcare: Deep Learning Based Mathematical Methods for Uncovering Large-Scale Forgery in Medical Images

Contact Info

Product

Resources

About