| Popis: |
As demonstrated in hereditary disorders and acute coronary syndrome, facial and physical signals (clinical gestalt) in Deep learning (DL) models enhance the evaluation of patients' health state. It is unknown whether adding clinical gestalt enhances the classification of patients with acute illnesses. The applicability of clinical gestalt may be assessed using simulated or augmented data, similar to earlier work on DL analysis of medical images.. In this study, using photos of facial cues for disease, For automatic rug sick identification, we developed a computer-aided diagnosis method. Individuals who were experiencing an acute sickness were seen by uninformed observers to have pale skin, lips and a more bloated face, more droopy eyelids, redder eyes, less shiny and spotted skin, as well as seeming more weary.. According to our research, critically ill and potentially contagious individuals can be identified using facial clues related to the skin, lips, and eyes. 1 To address the lack of data, we used deep transfer learning and constructed a CNN framework using the four transfers learning techniques shown below.: ResNet50, InceptionV3, VGG16, VGG19, Xception, and Inception. Whereas ResNet101 is utilized in the current methods, it does not have the appropriate precision and could use improvement. So, it is suggested to combine the current method with additional transfer learning techniques. The suggested method was examined using a publicly accessible dataset called Facial Cue of Illness. |
