TSRNet: Tongue image segmentation with global and local refinement

Cai, Wenjun; Zhang, Mengjian; Wen, Guihua; Yang, Pei

doi:10.1016/j.displa.2023.102601

Cited by 4 publications

(1 citation statement)

References 35 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, existing image processing methods still exhibit numerous deficiencies and limitations when confronted with the complex scenarios in critical care. On one hand, current technologies often struggle with the variability of patient conditions and the uncertainty of image quality in the segmentation of physiological state images, leading to insufficient accuracy and robustness [14][15][16]. On the other hand, in the classification of disease levels based on physiological signal images, existing models lack specificity and differentiation in feature expression, indicating that there is room for improvement in classification performance [17][18][19][20].…”

Section: Introductionmentioning

confidence: 99%

Automated Physiological Status Detection and Disease Evaluation of Critically Ill Patients via Image Processing Technologies

Wang,

Ma,

Zhao

et al. 2024

View full text Add to dashboard Cite

In the realm of critical care, precise detection of physiological status and disease evaluation are paramount for effective treatment and nursing. With the continuous advancement of medical imaging technology, image processing techniques herald new possibilities for achieving these objectives. This study is dedicated to enhancing the automation level and accuracy of physiological status monitoring and disease evaluation for critical patients through cutting-edge image analysis technologies. The background section explores the current application of medical imaging in critical care, underscoring the significance and developmental trends of automated image processing in this domain. The state-of-the-art review highlights existing image segmentation and classification methods, addressing challenges encountered in complex critical care scenarios, such as insufficient segmentation precision and weak feature representation capabilities. To tackle these issues, a novel image segmentation approach based on boundary learning and enhancement (BLE), along with a disease severity classification model leveraging feature augmentation, is proposed. Through optimization of deep learning models, the segmentation part strengthens the identification of subtle boundaries in images depicting the physiological status of critical patients, thereby enhancing segmentation accuracy and robustness. In the aspect of disease classification, the study improves the model's ability to recognize features indicative of the patients' condition through feature enhancement techniques, leading to heightened classification precision. The application of these methodologies not only elevates the quality of care but also aids healthcare professionals in making more rapid and accurate decisions. The outcomes of this study hold significant implications for advancing the level of automation in physiological status monitoring and disease evaluation of critical patients, and they positively impact the further development of medical imaging technology in clinical applications.

show abstract