CRV #**** CRV 2006 Submission ****. CONFIDENTIAL REVIEW COPY. DO NOT DISTRIBUTE. CRV #**** 1 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 Abstract This paper presents a GET (Generic Edge Token) based approach of detecting and recognizing objects by their shapes, and applies it to improve our ongoing work that considers ways of interacting with paper maps using a handheld. In our work, the GET-based technique aims to help user better locate points of interest on map by recognizing these icons from images/videos captured by handheld camera. In this method, video/image content can be described using a set of perceptual shape features called GETs. Perceptible object can be extracted from a GET map, and then be compared against pre-defined icon models based on GET shape features in recognition. This method provides a simple, efficient way to locate points of interest on the map, determining handheld location, orientation when combined with RFID (senor-based) technique. The tests show that the GET-based object identification can be executed in reasonable time for the real-time interaction system. Meanwhile, the detections and recognitions are robust under different lighting conditions, camera focus, camera rotation, and distance from the map. Keywords: Generic Edge Token, object recognition, ubiquitous interaction 1. Introduction Visio-based techniques allow tracking and recognizing visual features from video sequences. These techniques can be widely used in many applications [1] [2]. In this paper we present a GET (Generic Edge Token) based approach of detecting and recognizing objects in video/image by their shapes, and apply it to our ongoing work that considers ways of interaction with paper maps and kiosks using the handheld. In our work, the integration of GET-based object identification and Radio Frequency Identification (RFID) sensor-based techniques permits direct interaction with the visual features of the resource using handheld. The key idea in the “interaction with map and kiosks” research is to link the virtual and real worlds by hovering over, pointing at, or gesturing toward physical maps or kiosks with a handheld device. The handheld device acts as a window on the virtual world, allowing the user to send queries and gather feedbacks. For example, the user can select an icon on a map by centering it in the visual field of a handheld camera [3]. The system will detect and recognize an icon based on its GET shape features, determine the handheld location, orientation to the map by using the combination of video recognition results and RFID data, and then send back information associated with the icon to user. In the icon detection, a perceptible object can be extracted from a GET map by grouping the approximate GETs into object contour closures. The detected object is then compared against pre-defined icon models based on GET shape features for icon recognition. Most object identification techniques require detect object first from a video/image and then recognize the object by its features. The features used commonly in identification include color, edge, optical flow and texture, etc. The most common object detection techniques either segment objects from a single frame/image (e.g., region segmentation) or employ frame difference information during detection (e.g., background subtraction,). Some object detection methods are reviewed in [4][5]. An extensive survey for object detection, representation and tracking is presented in [6]. Comparing to other object identification techniques, our GET-based method uses the unique perceptually stable edge features to describe objects and represent video/image content. It provides a simple and efficient way to identify the icons in real-time under different conditions such as lighting, camera distance, rotation and shooting angle. The rest of this paper is organized as follows. Section 1 introduces backgrounds and motivation of the technique first and then gives system architecture. Section 2 presents the GET-based object detection. Section 3 provides object feature descriptors that can describe the detected objects and pre-defined icon models based on their shape features. Section 4 discusses the details of the icon recognition. In Section 5, experimental results and analysis are provided. Finally, Section 6 discusses conclusions and suggests possible future works. GET-based map icon Identification for Interaction with Map and Kiosks Huiqiong Chen, Derek Reilly Faculty of Computer Science, Dalhousie University, Halifax, Canada Paper ID ****< replace **** here and in header with paperID>