Deep Learning Algorithms for Detection and Classification of Gastrointestinal Diseases

Table 4

Comparison of the performance of our proposed system with models of previous studies.

Previous studies	Accuracy (%)	Sensitivity (%)	Specificity (%)	AUC (%)
Godkhindi and Gowda [27]	88.56	88.77	87.35	—
Pozdeev et al. [29]	88.00	93.00	82.00	—
Bour et al. [50]	87.10	87.10	93.00	—
Zhang et al. [51]	85.90	87.60	—	86.00
Ribeiro et al. [31]	90.96	95.16	74.19	—
Min et al. [32]	78.40	83.30	70.10	—
Zhu et al. [52]	85.00	83.00	82.54	—
Fonollá et al. [53]	90.20	90.10	90.30	97.00
Zhang et al. [51]	70.40	70.40	70.90	90.00
Proposed model (AlexNet)	97.00	96.80	99.20	99.98
Proposed model (GoogleNet)	96.70	96.60	99.00	99.99
Proposed model (ResNet-50)	95.00	94.80	98.80	99.69