Contour-Point Refined Mask Prediction for Single-Stage Instance Segmentation

Fig. 6. Network structures of different mask prediction subnetworks. (a) Using different prediction subnetworks (DPS); (b) using the same prediction subnetwork (SPS)

Download full size

Fig. 7. Segmentation results of each stage. (a) Original image; (b) segmentation results of semantic segmentation subnetwork; (c) segmentation results of mask prediction subnetwork; (d) final instance segmentation results

Download full size

Fig. 8. Instance segmentation of different methods. (a) Original image; (b) PolarMask; (c) our method

Download full size

Fig. 9. Results of the proposed algorithm under the MS COCO test dataset

Download full size

Table 1. Comparison of semantic segmentation feature fusion methodsunit: %
View table
Table 1. Comparison of semantic segmentation feature fusion methodsunit: %
Method AP AP₅₀ AP₇₅ AP_S AP_M AP_L
Concat+1×1 conv 30.5 51.0 31.9 12.5 32.7 46.3
Sum 30.8 51.5 32.3 12.4 32.0 45.2

Table 2. Comparison of experimental results of different mask prediction subnetworksunit: %
View table
Table 2. Comparison of experimental results of different mask prediction subnetworksunit: %
Method Figure AP AP₅₀ AP₇₅ AP_S AP_M AP_L
DPS Fig. 6(a) 29.4 49.8 29.9 12.3 31.6 43.2
SPS Fig. 6(b) 30.8 51.5 32.3 12.4 32.0 45.2

Table 3. Experimental results obtained by different loss function weightsunit: %
View table
Table 3. Experimental results obtained by different loss function weightsunit: %
λ_segm λ_angle AP AP₅₀ AP₇₅
1.0 1.0 30.2 50.7 31.8
0.5 1.0 30.0 50.5 31.8
1.0 0.5 30.5 51.2 32.0
1.0 0.3 30.6 51.3 32.2
1.0 0.2 30.8 51.5 32.3
1.0 0.1 30.7 51.3 32.3

Table 4. Comparison of each module under MS COCO-validation datasetunit: %

View table

Table 4. Comparison of each module under MS COCO-validation datasetunit: %

Method	AP	AP₅₀	AP₇₅	AP_S	AP_M	AP_L
Baseline	29.1	49.5	29.7	12.6	31.8	42.3
Baseline+Semantic segmentation subnetwork	29.1	50.6	30.1	11.7	30.6	44.7
Baseline+Mask prediction subnetwork	29.4	50.5	29.9	12.8	31.8	43.3
Ours	30.8	51.5	32.3	12.4	32.0	45.2

Table 5. Performance comparison of different methods under the MS COCO test datasetunit: %

View table

Table 5. Performance comparison of different methods under the MS COCO test datasetunit: %

Method	Backbone	AP	AP₅₀	AP₇₅	AP_S	AP_M	AP_L
MNC^[22]	Resnet101	24.6	44.3	24.8	4.7	25.9	43.6
FCIS^[5]	Resnet101	29.2	49.5	-	7.1	31.3	50.0
YOLACT^[12]	Resnet101	31.2	50.6	32.8	12.1	33.3	47.1
PolarMask^[16]	Resnet101	30.4	51.9	31.0	13.4	32.4	42.8
Ours	Resnet101	32.5	53.6	34.3	13.1	34.3	48.0

Tools

Get Citation

Copy Citation Text

Xuyi Zhang, Jiale Cao. Contour-Point Refined Mask Prediction for Single-Stage Instance Segmentation[J]. Acta Optica Sinica, 2020, 40(21): 2115001

Download Citation

EndNote(RIS)BibTex Plain Text

Set citation alerts for article

Save article for my favorites

Paper Information

Category: Machine Vision

Received: Jun. 8, 2020

Accepted: Jul. 15, 2020

Published Online: Oct. 26, 2020

The Author Email: Xuyi Zhang (zxy1996@tju.edu.cn)

DOI:10.3788/AOS202040.2115001

Topics

laser devices and laser physics

Lasers and Laser Optics

Laser physics

laser manufacturing

Instrumentation, Measurement and Metrology

Table 1. Comparison of semantic segmentation feature fusion methodsunit: %

Table 1. Comparison of semantic segmentation feature fusion methodsunit: %

Table 2. Comparison of experimental results of different mask prediction subnetworksunit: %

Table 2. Comparison of experimental results of different mask prediction subnetworksunit: %

Table 3. Experimental results obtained by different loss function weightsunit: %

Table 3. Experimental results obtained by different loss function weightsunit: %

Table 4. Comparison of each module under MS COCO-validation datasetunit: %

Table 4. Comparison of each module under MS COCO-validation datasetunit: %

Table 5. Performance comparison of different methods under the MS COCO test datasetunit: %

Table 5. Performance comparison of different methods under the MS COCO test datasetunit: %