Enhancing registration accuracy and eminence of multispectral transmission breast images by fusing multi-wavelength gen using vision transformer and LSTM.

Journal: Scientific reports
PMID:

Abstract

Enduring studies in the field of early breast cancer screening are investigating the use of multispectral transmission imaging. The frame accumulation system handles multispectral transmission images with deprived grayscale and unsatisfactory resolution. Nevertheless, the accuracy of the results may be compromised due to the motion caused by human breathing and camera instability during the transition between image frames. This study proposes a novel deep learning approach that combines a Vision Transformer (ViT) with Long Short-Term Memory (LSTM) networks to enhance image registration accuracy and reduce noise in multispectral transmission images. The registration accuracy and image quality of frame accumulation after registration are assessed distinctly. The results are stated using the metrics for each wavelength image as Mutual Information (MI), Peak Signal-to-Noise Ratio (PSNR), Coefficient Correlation (CC), Root Mean Square Error (MSE), and Registration Time (RT). The suggested technique for early breast cancer screening image registration and denoising is endorsed using multispectral transmission images at 600, 620, 670, and 760 nm wavelengths. The statistical assessment of the outcomes offers support for the endorsed approach. In this study, the Vision Transformer and LSTM fused approach demonstrates significantly improved accuracy, registration time, and preservation of the registered and grayscale images. It establishes the foundation for early detection of breast cancer and anomalies through enhanced image quality. The aggregated metrics comprehensively assess the performance of the whole image dataset.

Authors

  • Muhammad Fahad
  • Tao Zhang
    Department of Traumatology, Chongqing Emergency Medical Center, Chongqing University Central Hospital, School of Medicine, Chongqing University, Chongqing, 40044, People's Republic of China.
  • Sajid Ullah Khan
    Universiti Malaysia Sarawak, Kota Samarahan, Sarawak, Malaysia.
  • Abdullah Albanyan
    Software Engineering Department, College of Computer Engineering and Sciences, Prince Sattam bin Abdulaziz University, Alkharj, Saudi Arabia.
  • Yanzhang Geng
    School of Electrical and Information Engineering, Tianjin University, 92 Weijin Road, Nankai District, Tianjin, 300072, China.
  • Xin Zhao
    Florida International University.
  • Nan Su Su Win
    Medical School of Tianjin University, Tianjin, 300072, China.
  • Gang Li
    The Centre for Cyber Resilience and Trust, Deakin University, Australia.
  • Ling Lin
    Center for Sleep Science and Medicine, Stanford University, Stanford, 94304, CA, USA.
  • Fazeela Siddiqui
    School of Electrical and Information Engineering, Tianjin University, 92 Weijin Road, Nankai District, Tianjin, 300072, China.