A novel network-level fused deep learning architecture with shallow neural network classifier for gastrointestinal cancer classification from wireless capsule endoscopy images

Table 8 Analysis of proposed model based on various variants

Kvasir-VI
Variants	Parameters	V(Accuracy)	Training Time
Baseline DensNet201	20 M	94.92	11 h 6 m
DensNet201 + SA	20.6 M	95.14	12 h 15 m
Proposed SC-DSAN	12.4 M	97.82	9 h 40 m
Proposed Network Level Fusion	14.7 M	99.19	10 h 21 m
Kvasir-VII
Base DensNet201	-	90.94	9 h 18 m
DensNet201 + SA	-	92.66	11 h 27 m
Proposed SC-DSAN	-	94.18	8 h 16 m
Proposed Network Level Fusion	-	95.24	9 h 48 m

ISSN: 1472-6947