Tuesday, 3 May 2022

Designing a High-performance Deep Learning Theoretical Model for Biomedical Image Segmentation by using Key Elements of the Latest U-Net-Based Architectures: A Recent Study| Chapter 9 | Research Developments in Science and Technology Vol. 2

Current advancements in machine learning, particularly deep learning, are proving to be useful in identifying and quantifying patterns in medical pictures. The crucial power of deep learning algorithms to generate hierarchical feature representations directly from pictures, which eliminates the need for handcrafted features, lies at the heart of these breakthroughs. Deep learning is quickly becoming the state-of-the-art for medical image processing, with performance gains in a wide range of clinical applications.


We want to develop a highly automated technique for identifying and staging precancerous and cervical cancers, as well as thyroid malignancies, that is improved using deep learning (DL) and confirmed by a randomised controlled clinical trial. We wish to use U-Net-based architectures to build a high-performance deep learning model for medical image segmentation that is independent of organs/tissues, dimensions, or image type (2D/3D) and verify it in a randomised, controlled clinical trial. Based on the U-Net-based architectures that we envisaged, we mostly used U-Net-based architecture analysis to identify the primary aspects that we considered were important in the design and optimization of the integrated DL model. Second, we'll test the DL model's performance in a randomised, controlled clinical study. Precancers, cervical cancer, and thyroid cancer will all be diagnosed and staged using our DL model, which will be highly automated. We created a composite model that takes into consideration the major elements of each architecture. Attention gate mechanism is an improvement added to convolutional network architecture for fast and precise image segmentation (Attention U-Net), Harmony Densely Connected Network-Medical image Segmentation, Overcomplete Convolutional Network Kite-Net (Kite-Net), Overcomplete Convolutional Network Kite-Net (Kite-Net), Overcomplete Convolutional Network Kite-Net (Kite-Net), Overcomplete Convolutional Network Kite-Net (Kite-Net (HarDNet-MSEG). In this context, we will create a comprehensive computer-assisted diagnostic method that will be verified in a randomised controlled trial. For precancers, cervical cancer, and thyroid cancer, the model will be a highly automated diagnostic and staging tool. This would save time and effort for professionals when analysing medical pictures, help build a better therapeutic approach, and give a "second opinion" on computer-assisted diagnosis.

Author(s) Details:

Andreea Roxana Luca,
Faculty of General Medicine, “Grigore T. Popa” University of Medicine and Pharmacy, Iasi, Romania and Departament Obstetrics and Gynecology, Integrated Ambulatory of Hospital Sf. Spiridon”, Iasi, Romania.

Tudor Florin Ursuleanu,
Faculty of General Medicine, “Grigore T. Popa” University of Medicine and Pharmacy, Iasi, Romania and Department of Surgery VI, “Sf. Spiridon” Hospital, Iasi, Romania and Departament of Surgery I, Regional Institute of Oncology, Iasi, Romania.

Liliana Gheorghe,
Faculty of General Medicine, “Grigore T. Popa” University of Medicine and Pharmacy, Iasi, Romania and Department of Radiology, “Sf. Spiridon” Hospital, Iasi, Romania.

Roxana Grigorovici,
Faculty of General Medicine, “Grigore T. Popa” University of Medicine and Pharmacy, Iasi, Romania.

Stefan Iancu,
Faculty of General Medicine, “Grigore T. Popa” University of Medicine and Pharmacy, Iasi, Romania.

Maria Hlusneac,
Faculty of General Medicine, “Grigore T. Popa” University of Medicine and Pharmacy, Iasi, Romania.

Cristina Preda,
Faculty of General Medicine, “Grigore T. Popa” University of Medicine and Pharmacy, Iasi, Romania and Department of Endocrinology, “Sf. Spiridon” Hospital, Iasi, Romania.

Alexandru Grigorovici,
Faculty of General Medicine, “Grigore T. Popa” University of Medicine and Pharmacy, Iasi, Romania and Department of Surgery VI, “Sf. Spiridon” Hospital, Iasi, Romania.

Please see the link here: https://stm.bookpi.org/RDST-V2/article/view/6611

No comments:

Post a Comment