A Novel Road Traffic Sign Detection and Recognition Approach by introducing CCM and LESH Usman Zakir 1 , Asima Usman, Amir Hussain 1 1 COSPIRA laboratory, Division of Computing Science, School of Natural Sciences University of Stirling, Stirling FK9 4LA usmanzakir@gmail.com, au@instantclaims.org.uk a.hussain@cs.stir.ac.uk Abstract. A real time road sign detection and recognition system can provide an additional level of driver assistance leading to an improved safety to passengers, road users and other vehicles. Such Advanced Driver Assistance Systems (ADAS) can be used to alert a driver about the presence of a road sign by reducing the risky situation during distraction, fatigue and in the presence of poor driving conditions. This paper is divided into two parts: Detection and Recognition. The detection part includes a novel Combined Colour Model (CCM) for the accurate and robust road sign colour segmentation from video stream. It is complemented by a novel approach to road sign recognition which is based on Local Energy based Shape Histogram (LESH). Experimental results and a detailed analysis to prove the effectiveness of the proposed vision system are provided. An accuracy rate of above 97.5% is recorded. Keywords: Colour Segmentation, Detection, Recognition, CCM, LESH, SVM, ADAS. 1 Introduction Road signs have meanings depending on their colours, shapes used and contents included within. Primarily, road sign colours are Red, Blue, Green, Brown, Yellow or White, which signify and categorize their importance, e.g. Red for obligatory signs and Blue for advisory signs. Therefore, colour plays an important initial role in a typical road sign detection task. Similarly, the global and local shape related features of a road sign can provide important clues in distinguishing one sign from another, i.e. in the recognition of the detected road signs. Due to varying lighting and weather conditions, segmentation of road signs using colour information and their recognition based on shape features; especially in outdoor images is a significantly challenging task. A detailed literature review carried out by the authors on automatic road sign detection and recognition revealed that even though a significant amount of research