. 2022 Oct 15;8(10):285. doi: 10.3390/jimaging8100285

Table 2.

Average precision (AP) results for both models for two different image input resolutions.

Model	S-VED (1280 × 1280)	S-VED(640 × 640)	S-VED $_{e x t e n d e d}$ (1280 × 1280)	S-VED $_{e x t e n d e d}$ (640 × 640)
YOLOv5	0.962	0.974	0.983	0.986
Faster-RCNN	0.949	0.940	0.975	0.965
DocExtractor	0.890	0.735	0.925	0.809