Matching Multi-Scale Features and Prediction Tasks for Real-Time Object Detection

Fig. 1. Existing object detection algorithms using convolution feature to complete prediction task. (a) CenterNet detection model; (b) detection model based on multi-scale feature for total task prediction; (c) proposed MFT detection model

Download full size

Fig. 2. The proposed MFT network structure

Download full size

Fig. 3. MSH module architecture

Download full size

Fig. 4. MRFH module architecture

Download full size

Fig. 5. Architectures of different feature fusion methods

Download full size

Fig. 6. Visual effect comparison of CenterNet and MFT detector on COCO dataset

Download full size

Fig. 7. Visual effect comparison of CenterNet and MFT detector on COCO dataset

Download full size

Table 1. Comparison of different object detection algorithms on the COCO dataset

View table

Table 1. Comparison of different object detection algorithms on the COCO dataset

Condition	Method	Backbone	Size	V /(frame·s^-1)	mAP /%	AP_S /%	AP_M /%	AP_L /%
V>60 frame/s	SSD^[8]	VGG16	300×300	60.6	23.2	5.3	23.2	39.6
	SSD^[8]	MobileNetV2	512×512	110.7	22.1	5.8	16.9	43.6
	CenterNet^[14]	Res18	512×512	128.5	28.1	10.1	31.5	42.6
	TTFNet^[30]	Res18	512×512	112.3	28.1	11.8	29.5	41.5
	MTF	Res18	512×512	94.5	31.5	14.9	35.3	44.3
V<60 frame/s	FCOS^[15]	Res18	1330×800	20.8	26.9	13.9	28.9	36
	CenterNet^[14]	Res101	512×512	45.1	34.6	10.1	31.5	42.6
	SSD^[8]	VGG16	512×512	23.4	26.8	9.0	28.9	41.9
	YOLOv3^[29]	D53	608×608	30.3	33.0	18.3	25.4	41.9
	EfficientDet^[28]	EfficientNet	512×512	47.1	33.8	12.4	34.7	54.4
	MTF	Res50	512×512	54.9	35.3	12.9	34.3	44.3

Table 2. Ablation results of different proposed modules on the PASCAL VOC dataset
View table
Table 2. Ablation results of different proposed modules on the PASCAL VOC dataset
Module ECF MSH MRFH LWS mAP /%
CenterNet (baseline) 70.64
+ECF √ 72.28
+MSH √ √ 73.16
+MRFH √ √ √ 73.86
MFT √ √ √ √ 74.09

Table 3. Reasonability of MSH module and MRFH module
View table
Table 3. Reasonability of MSH module and MRFH module
Module MSH MRFH MSH-mismatch MRFH-mismatch mAP /%
All-mismatch √ √ √ √ 73.01
MSH-mismatch √ √ √ 73.18
MFT √ √ 73.86

Table 4. Complementary experiment of MRFH and MSH
View table
Table 4. Complementary experiment of MRFH and MSH
MSH MRFH mAP /% ΔmAP
× × 72.28 0
√ × 73.16 +0.88
× √ 73.44 +1.16
√ √ 73.86 +1.58

Table 5. Ablation results of different scale features reused by MSH module
View table
Table 5. Ablation results of different scale features reused by MSH module
Module Large Medium Small mAP/%
MSH-L √ 73.44
MSH-LM √ √ 73.51
MSH √ √ √ 73.86

Table 6. Experiment of different feature fusion methods
View table
Table 6. Experiment of different feature fusion methods
Fusion method w₁ w₂ w₃ w₄ mAP /%
Simple average 1/4 1/4 1/4 1/4 73.86
Learned weight 0.2894 0.2551 0.3269 0.3307 74.09

Tools

Get Citation

Copy Citation Text

Hongjie Du, Hanqing Sun, Jiale Cao, Yanwei Pang. Matching Multi-Scale Features and Prediction Tasks for Real-Time Object Detection[J]. Laser & Optoelectronics Progress, 2021, 58(12): 1210014

Download Citation

EndNote(RIS)BibTex Plain Text

Set citation alerts for article

Save article for my favorites