Multiscale Feature Learning Based on Enhanced Feature Pyramid for Vehicle Detection

Table 3

Comparative results of GFLOPs with different models.

Model	Input resolution	GFLOPs
YOLOv2 [3]	416 × 416	17.4
YOLOv3 [40]	416 × 416	62.8
Faster R-CNN [4]	∼600 × 1000	57.9
Proposed model	∼600 × 1000	73.2