| Abstrakt: |
An innovative model for TB detection is proposed using chef leader-based optimization (CLBO)-based Dense convolutional network (DenseNet). Here, pre-processing is performed using an adaptive bilateral filter, whereas segmentation is carried out using ResUNet+ +. At last, TB detection is accomplished through DenseNet. Moreover, the networks ResUNet ++ and DenseNet are trained by the proposed CLBO, which is an integration of the chef-based optimization algorithm (CBOA) and hybrid leader-based optimization (HLBO). Additionally, the modeled CLBO_DenseNet has outperformed other classical models by delivering accuracy, TPR, TNR, PPV, NPV, F1-score, and dice coefficient with the values of 98.90%, 97.60%, 94.60%, 94.60%, 93.30%, 96.10%, and 96.50%, respectively. [ABSTRACT FROM AUTHOR] |
Full text from SpringerLink