Minireviews
Copyright ©The Author(s) 2025.
Artif Intell Gastrointest Endosc. Sep 8, 2025; 6(3): 108281
Published online Sep 8, 2025. doi: 10.37126/aige.v6.i3.108281
Table 1 Comparison of two machine learning models

Supervised model
Unsupervised model
Date dependencyHigh-quality annotationsLabel-efficient
Clinical applicabilityTask-specific optimizationExploratory applications
InterpretabilityModerate (via feature mapping)Low (abstract feature hierarchy)
DeploymentDevice-specific optimizationGeneralizable frameworks
Table 2 Comparison of the effects of machine learning-based endoscopic image recognition in different diseases

Sensitivity (range)
Specificity (range)
Key influencing factors
Barrett’s esophagus85%-92%78%-88%Complexity mucosal texture, annotation consistency
Early gastric cancer89%-95%82%-91%Lesion size, endoscopic image resolution
Ulcerative colitis76%-84%88%-93%Subjectivity of inflammation activity scoring, data heterogeneity
Colorectal polyps92%-97%90%-95%Polyp morphological diversity, real-time detection algorithm efficiency