ExpACVO-Exponential Anti Corona Virus Optimization Enabled Hybrid Deep learning for Tongue Image Segmentation Towards Diabetes Mellitus Detection Via CNN-LSTM Network

The surprising headways in biotechnology and public medical services foundations have prompted a pivotal creation of basic and touchy medical care information. By applying canny information examination methods, many fascinating examples are distinguished for the early and beginning discovery and avoidance of a few deadly infections. Diabetes Mellitus (DM) is a very dangerous illness since it adds to other deadly infections, i.e., heart, kidney, and nerve harm. The following stages have been proposed in this paper for the order, beginning phase ID, and expectation of diabetes: a) Data collection from tongue dataset of many images, b) Preprocessing using the gaussian filter for noise removal, c) Segmentation using U-Net-CRFRNN for segmenting region of interest and finally d) classification using the integration of Exponential Weighted Moving Average (EWMA) and Anti Corona Virus Optimization (ACVO) and also to increase the efficiency of the network, use of CNN-LSTM. The suggested architecture is more effective than other state-of-the-art models on many different parameters, based on the outcomes (accuracy: 0.94, sensitivity:
$0.96$
, specificity: 0.96)