D. Lopresti, J. Hu, and R. Kashi (Eds.): DAS 2002, LNCS 2423, pp. 66–69, 2002. © Springer-Verlag Berlin Heidelberg 2002 Automatic Reading of Traffic Tickets Nabeel Murshed Intelligent Information Systems A Pattern Recognition and Information Technology Company Dubai, United Arab Emirates Abstract. The present work presents a prototype system to extract and recog- nize handwritten information in a traffic ticket, and thereafter feeds them into a database of registered cars for further processing. Each extracted information consists either of handwritten isolated Arabic digits or tick mark “x”. The ticket form is designed in such a way to facilitate the extraction process. For each in- put, the output of the recognition module is a probabilistic value that indicates the system confidence of the correct pattern class. If the probabilistic output is less than the determined threshold, the system requests assistance from the user to identify the input pattern. This feature is necessary in order to avoid feeding in wrong information to the database, such as associating the traffic ticket with the wrong registered car. 1. Introduction Automatic reading of documents has become an interesting application area of Document Analysis Systems. Many systems have been developed for a wide range of applications. The reader is referred to the proceedings of the IAPR Workshops on Document Analysis Systems [1–4]. Most systems have been targeted towards Latin, Germanic, and Far Eastern languages. The present work is aimed at extracting and recognizing Arabic handwritten numerals for police application. To our knowledge, no system have been developed so far for such application. Independent of the language, most automatic reading systems share two principle requirements: high recognition rate and short processing time, which are directly re- lated to each other. In almost all cases, achieving high recognition rates could yield a relatively long processing time. One may argue that with the high speed of personal computers, one could obtain acceptable recognition results with acceptable processing time. A third important requirement is the degree of human interaction to increase the recognition rate, particularly in critical applications. The objective of the proposed system is to extract and recognize six pieces of handwritten information from a traffic violation ticket. The information represent the following: date and time of violation, number and color of the number plate, violation type, and the policeman’s ID number. All of those information are handwritten nu- merals except the color and violation type which are tick mark “x”. The recognized information are put into a record associated with the number plate. Figure 1, in page 4, shows an example of the traffic violation form designed for the application at hand. The paper is organized as follow. Section 2 and 3 describes, respectively, system architecture and experimental results. Comments and conclusion are given in section 3.