°h area 
ıterference 
ental area 
ines, trees, 
milar with 
ippearance 
Compared 
shows our 
d suppress 
cy map 
d 
1999 and 
n signs etc. 
three kinds 
prohibition 
hod is also 
e than 640 
ustrate the 
ie. original 
he process 
  
3) Detection of circular Signs 
The experimental results show the method has high accuracy in 
locating the circular traffic signs. Some experimental results are shown 
in Fig.6(a)~(d). Fig6(d) shows the circular indication sign was 
detected by this method. These results show that the method is robust 
and scalable. 
  
  
(c) (d) 
Fig 6. Experimental Images 
We use 673 street view images to verify the performance of the 
proposed method, which contains 261 traffic signs. Compared results 
of different methods are shown in tab 1, our method achieves the high 
detection rate than other saliency detection method, but the high false 
alarm rate is a problem need to be concerned. 
Tab 1. Compared results of different methods 
  
  
  
  
Mae oen m | Rue. 
Itti 161 162 62% 
Hou 116 212 44% 
Our method 217 206 83% 
  
  
  
  
  
  
4. .CONCLUSIONS 
In this work, we proposed a new detection method of traffic sign 
according to the design principle of the traffic sign. A visual attention 
model based on two-way integration is used to analyze the whole scene 
to acquire the candidate area. And then these candidate areas will be 
analyzed according to the shape characteristics of the traffic sign to 
detect traffic signs. In addition to, based on the mechanism of 
biological vision, we describe a saliency map generation method 
based on visual contrast and apply this method into the bottom-up 
phase in visual attention model. Experimental results show our 
method has a high detection probability than other saliency 
detection method. In the future, we will integrate the visual attention 
model with the ability of learning and memory to realize the traffic sign 
detection model fusioned with visual attention, learning and memory. 
ACKNOWLEDGEMENTS 
This work was supported by NSFC (40971232,41023001), 
National 
Support Program (2012BAH35B03), 863 Program 
(2011AA010500). 
[1] 
[2] 
[4] 
[5] 
[6] 
[7] 
[8] 
[10] 
[11] 
[12] 
[13] 
REFERENCE 
Houben S.A single target voting scheme for traffic 
sign detection[C].Intelligent Vehicles Symposium 
(IV), 2011 IEEE, 2011 
Kyung-in M,Jae-Saek O, Byong-Woo K.Traffic sign 
extract and recognition on unmanned vehicle using 
image processing based on support vector 
machine[C].Control, Automation and Systems 
(ICCAS), 2011 11th International Conference on, 
2011 
Rajesh R,Rajeev K,Suchithra K, et al.Coherence 
vector of Oriented Gradients for traffic sign 
recognition using Neural Networks[C].Neural 
Networks (IJCNN), The 2011 International Joint 
Conference on, 2011 
Sallah S S M,Hussin F A, Yusoff M Z.Road sign 
detection and recognition system for real-time 
embedded applications[C].Electrical, Control and 
Computer Engineering (INECCE), 2011 International 
Conference on, 2011 
Xiaoguang H,Xinyan Z,Hui L, et al Traffic sign 
detection based on visual attention 
model[C].Geoinformatics, 2011 19th International 
Conference on, 2011 
Yanhua J,Shengyan Z,Yan J, et al Traffic sign 
recognition using Ridge Regression and OTSU 
method[C].Intelligent "Vehicles Symposium (IV), 
2011 IEEE, 2011 
Zhixie CJing Y, Bin K.A Robust Traffic Sign 
Recognition System for Intelligent Vehicles[C].Image 
and Graphics (ICIG), 2011 Sixth International 
Conference on, 2011 
Hazelhoff L,Creusen I, de With P H N.Robust 
detection, classification and positioning of traffic 
signs from street-level panoramic images for 
inventory purposes[C].Applications of Computer 
Vision (WACV), 2012 IEEE Workshop on, 2012 
Jung-Guk P, Kyung-Joong K.A method for feature 
extraction of traffic sign detection and the system for 
real world scene[C].Emerging Signal Processing 
Applications (ESPA), 2012 IEEE International 
Conference on, 2012 
Liu Y,Ki-Ryong K,KwangSeok M, et al.Broken 
traffic sign recognition based on local histogram 
matching[C ].Computing, Communications and 
Applications Conference (ComComAp), 2012, 2012 
Iti L,Gold C, Koch C.Visual attention and target 
detection in cluttered natural scenes [J]. Optical 
Engineering, 2001, 40: 1784 
Asakura T,Aoyagi Y, Hirose O K.Real-time 
recognition of road traffic sign in moving scene image 
using new image filter[C].SICE 2000. Proceedings of 
the 39th SICE Annual Conference. International 
Session Papers, 2000 
Iti L,Koch C, Niebur E.A model of saliency-based 
visual attention for rapid scene analysis [J]. Pattern 
Analysis and Machine Intelligence, IEEE 
Transactions on, 1998, 20(11): 1254-1259