Bibliometric Analysis of Image Segmentation with Deep Learning: An Analytical Study

This study presents a comprehensive bibliometric analysis of research pertaining to the utilization of deep learning techniques for image segmentation using CNN algorithms. A dataset comprising 1078 publications from journals and conference proceedings between 2018 and 2022 was examined through keyword search, co-occurrence network analysis, and keyphrase analysis. The study offers valuable insights into the present research landscape and identifies prominent areas of investigation, encompassing deep learning's application in medical imaging, disease detection, and object detection. Our analysis reveals China as a leading contributor to deep learning research for image segmentation, with 279 publications and 4,380 citations in Scopus. Advances in Biomedical Optics and Imaging—Proceedings of SPIE are pinpointed as the most productive source for deep learning research in image segmentation, with a specific emphasis on biomedical optics and imaging. The co-occurrence network analysis highlights the red cluster primarily focusing on methods and algorithms associated with deep learning in image segmentation. In contrast, the blue cluster demonstrates the application of these methods and algorithms to other objects or methods, with the “human” node standing out in terms of frequency and centrality. Our keyphrase analysis reveals the growing trend of Cellular Artificial Neural Networks over the past five years, indicating a shift in research focus toward this domain. Overall, this study's findings demonstrate the potential of deep learning to deliver precise and efficient segmentation of medical images, thereby enhancing clinical outcomes and patient care. Furthermore, our study contributes to an enhanced understanding of the current research landscape and identifies avenues for future exploration in deep learning techniques for image segmentation.