Lightweight deep learning model for gastrointestinal precancerous lesion screening with attention enhancement.

While the improvement in polyp recall did not reach statistical significance ( p = 1.000, McNemar’s exact test) likely due to the limited size of the test set, the SCA module achieved a clinically meaningful absolute improvement in polyp precision (99.04–100.00%), which is crucial for reducing false positives in clinical screening scenarios.

also in 111,027 other papers