Unveiling Precision in Abdominal Organ Segmentation: A Deep Dive into Emerging Deep Learning Paradigms for Single and Multi-modal Image Segmentation

Medical imaging is essential for diagnosing and managing diseases that impact human organs. Two widely used imaging techniques, Computed Tomography (CT) and Magnetic Resonance Imaging (MRI), offer valuable information about the structural aspects of organs. However, relying solely on a single imaging modality can sometimes limit disease detection accuracy. To address this limitation, the integration of multiple modalities followed by segmentation has gained traction, offering improved precision in assessing organ health. Accurate segmentation of organs from multi-modal medical images forms the cornerstone of modern healthcare, facilitating precise treatment planning, early disease detection, and personalized medicine. This paper offers an in-depth review of the latest trends and challenges in abdominal organ segmentation using deep learning approaches. It explores the use of attention mechanisms, Graph Neural Networks (GNNs) alongside traditional methods such as Convolutional Neural Networks (CNN), Fully Convolutional Networks (FCN), and Generative Adversarial Networks (GAN) within the scope of abdominal organ segmentation. The paper also addresses key challenges and opportunities in the field, highlighting the importance of continued innovation and collaboration to advance abdominal organ segmentation for improved clinical outcomes