Table 1.
Features and challenges on existing systems.
Fig 1.
Framework of MSLDSSC model.
Fig 2.
Modified Segnet framework.
Fig 3.
Standard form of SqueezeNet model.
Fig 4.
Structural representation of SDPA-SqueezeNet model.
Fig 5.
Image Results for Lung Disease Segmentation and Severity Classification using CT Images a) Input Images and b) Improved Weiner Filtering.
Table 2.
Training and testing images for HRCTCov19 dataset.
Fig 6.
Image Results for Lung Disease Segmentation using CT images a) Input Image b) FCM c) nnU-Net d) K-Means e) Conventional SegNet and f) Modified SegNet.
Table 3.
Segmentation analysis on modified SegNet over conventional approaches.
Table 4.
Statistical analysis for segmentation approaches.
Fig 7.
Positive metric analysis on SDPA-SqueezeNet+DCNN and conventional methods.
Fig 8.
Negative metric analysis on SDPA-SqueezeNet+DCNN and conventional methods.
Fig 9.
Other metric analysis on SDPA-SqueezeNet+DCNN and conventional methods.
Table 5.
Ablation evaluation on SDPA-SqueezeNet+DCNN over modified SegNet without Wiener filter, SqueezeNet and DCNN with images directly as input, Images + Wiener filter as input, Images + Wiener filter + segmentation, model with Wiener filter, model with conventional Segnet and model with MLDN.
Table 6.
Statistical analysis of accuracy.
Fig 10.
Analysis on Confusion Matrix a) F-RNN-LSTM [28] b) MCCLLD-CNN [29] c) CNN d) LinkNet e) LSTM f) SqueezeNet g) SVM h) ResNet and i) SDPA-SqueezeNet+DCNN.