Low-Parameter Real-Time Image Segmentation Algorithm Based on Convolutional Neural Network

Guanghong Tan; Jin Hou; Yanpeng Han; Shuo Luo

doi:10.3788/LOP56.091003

Journals >Laser & Optoelectronics Progress >Volume 56 >Issue 9 >Page 091003 > Article

Laser & Optoelectronics Progress
Vol. 56, Issue 9, 091003 (2019)

Low-Parameter Real-Time Image Segmentation Algorithm Based on Convolutional Neural Network

Guanghong Tan, Jin Hou^*, Yanpeng Han, and Shuo Luo

Author Affiliations

School of Information Science and Technology, Southwest Jiaotong University, Chengdu, Sichuan 611756, China

show less

DOI: 10.3788/LOP56.091003 Cite this Article Set citation alerts

Guanghong Tan, Jin Hou, Yanpeng Han, Shuo Luo. Low-Parameter Real-Time Image Segmentation Algorithm Based on Convolutional Neural Network[J]. Laser & Optoelectronics Progress, 2019, 56(9): 091003 Copy Citation Text

show less

Fig. 1. Convolution kernel. (a) Classical convolution kernel; (b) dilated convolution kernel Rrate=2; (c) dilated convolution kernel Rrate=3

Download full size

Fig. 2. Atrous-Fire modular structure

Download full size

Fig. 3. Dilated convolution kernel and initial characteristic graphs. (a) Sawtooth structure convolution kernel; (b) no grid feature graph; (c) grid feature graph

Download full size

Fig. 4. Network structure of Atrous-squeezeseg

Download full size

Fig. 5. Training loss value curves

Download full size

Fig. 6. Validation loss value curves

Download full size

Fig. 7. Effect comparison of ADE20K. (a) Original images; (b) ground truth; (c) proposed algorithm; (d) Squeezeseg+FCN; (e) VGG16+FCN; (f) SqueezeNet+FCN; (g) without dilated; (h) without BN

Download full size

Fig. 8. Effect comparison of PASCAL VOC. (a) Original images; (b) ground truth; (c) proposed algorithm; (d) Squeezeseg+FCN; (e) VGG16+FCN; (f) SqueezeNet+FCN; (g) without dilated; (h) without BN

Download full size

Layer name	Output size	Squeeze(S1)	Expand(E1/E3)	R_rate
Input image	224×224×3
Conv1	112×112×64
Maxpool1	56×56×64
Atrous-Fire1 (3×)	56×56×256	16	32	2/5/7
Maxpool2	28×28×256
Atrous-Fire2 (3×)	28×28×256	32	64	2/3/5
Maxpool3	14×14×256
Atrous-Fire3 (3×)	14×14×256	64	128	2/3/5
Atrous-Fire4 (2×)	14×14×512	128	256	1/2
Atrous-Fire4 (2×)	14×14×512	128	256	1/1

Table 1. Encoder parameters

Method	Number ofparameters	M_IU	Building	Sky	Car	Tree	Road	Person	Floor	Wall
Atrous-squeezeseg	21.09	62.9	67.5	84.0	61.4	58.1	64.7	49.1	60.4	58.5
Squeezeseg+FCN	54.65	55.9	61.8	85.8	48.4	51.8	61.5	32.3	53.8	52.2
VGG16+FCN	66.21	63.2	68.3	86.8	61.1	58.2	66.0	48.5	58.3	57.4
SqueezeNet+FCN	54.65	50.5	46.7	83.8	44.7	51.8	55.5	28.0	47.3	46.7
Atrous-squeezeseg(without dilated)	21.09	50.6	51.1	83.5	41.2	43.8	53.8	29.7	51.4	50.1
Atrous-squeezeseg(without BN)	21.09	51.6	51.3	83.5	43.6	45.8	58.0	29.8	50.1	51.3

Table 2. Number of parameters of different semantic segmentation models and MIU

Method	FPS /(frame·s^-1)		P_A /%
Method		GTX 1080Ti	NVIDIA TX2
Atrous-squeezeseg	45.3	8.3	59.5
Squeezeseg+FCN	39.5	4.2	59.3
VGG16+FCN	29.6	1.9	59.8
SqueezeNet+FCN	46.6	4.5	55.6
Atrous-squeezeseg(without dilated)	45.6	8.4	56.1
Atrous-squeezeseg(without BN)	56.2	9.2	57.3

Table 3. PA and FPS of model in different devices

Guanghong Tan, Jin Hou, Yanpeng Han, Shuo Luo. Low-Parameter Real-Time Image Segmentation Algorithm Based on Convolutional Neural Network[J]. Laser & Optoelectronics Progress, 2019, 56(9): 091003

Download Citation

Set citation alerts for the article

Tools

Set citation alerts for the article

Save the article for my favorites

Paper Information