Object Detection Algorithm Based on Improved Feature Pyramid

Table 5

COCO test 2017-dev detection results.

	Backbone	AP	AP0.5	AP0.75	APS	APM	AP L
Faster R-CNN	ResNet-101	34.9	55.7	37.4	15.6	38.7	50.9
Faster R-CNN	ResNet-101-FPN	36.2	59.1	39.0	18.2	39.0	48.2
Faster R-CNN^∗	ResNet-50-FPN	36.2	58.5	38.9	21.0	38.9	45.3
Faster R-CNN^∗	ResNet-101-FPN	38.8	60.9	42.1	22.6	42.4	48.5
Faster R-CNN^∗	ResNeXt-101(32 × 4d)-FPN	40.4	62.2	43.6	24.0	43.8	50.3
Mask R-CNN	ResNet-101-FPN	29.9	60.3	41.7	20.1	41.1	50.2
R-FCN	ResNet-101	33.1	51.9	—	10.8	32.8	45.0
CoupleNet	ResNet-101	37.7	53.5	35.4	11.6	36.3	50.1
YOLO v2	Darknet-19	21.6	44.0	19.0	5.2	22.4	35.5
YOLO v3	Darknet-53	33.0	57.9	34.4	18.3	35.4	41.9
SSD512	VGG-16	31.2	50.4	33.3	10.2	34.5	49.8
DSSD513	ResNet-101	33.2	53.3	35.2	13.0	35.4	51.1
RetinaNet	ResNet-101-FPN	39.1	59.1	42.3	21.8	42.7	50.2
RFB Net512	VGG-16	33.8	54.2	35.9	16.2	37.1	47.4
M2Det [29]	VGG-16	33.5	53.4	35.6	14.4	37.6	47.6
IBi-FPN [11]	VGG-16	32	50.9	33.7	12	37	47.1
M2Det [29]	ResNet-101	38.8	59.4	41.7	20.5	43.9	53.4
NETNet [30]	ResNet-101	38.5	58.6	41.3	19.0	42.3	53.9
MREFP-Net [10]	ResNet-101	39.3	60.1	43.1	20.6	43.9	52.0
MREFP-Net [10]	VGG-16	38.5	59.1	41.1	18.6	42.9	49.4
Ours	ResNet-50-AFPN	39.5	60.3	40.5	22.2	40.8	46.5
Ours	ResNet-101-AFPN	40.6	62.1	42.9	23.1	43.0	49.5
Ours	ResNeXt-101(32 × 4d)-AFPN	41.8	64.1	44.1	24.1	44.2	51.4