Identifying people in camera networks using wearable accelerometers

We propose a system to identify people in a sensor network. The system fuses motion information measured from wearable accelerometer nodes with motion traces of each person detected by a camera node. This allows people to be uniquely identified with the IDs the accelerometer-node that they wear, while their positions are measured using the cameras. The system can run in real time, with high precision and recall results. A prototype implementation using iMote2s with camera boards and wearable TI EZ430 nodes with accelerometer sensorboards is also described. Categories and Subject Descriptors C.3 [Special-Purpose and Application-Based Systems]: Real-time and embedded systems; I.4 [Image Processing and Computer Vision]: Scene Analysis—Sensor fusion, Tracking General Terms Measurement, Design, Experimentation Keywords Unique identification, Consistent labelling, Association problem

• Type: Conference
• Year: 2009
• Where: PETRA
• Authors: Thiago Teixeira, Deokwoo Jung, Gershon Dublon, Andreas Savvides
• Keywords: Accelerometer, Emerging Technology, Fuses Motion Information, PETRA 2009, Wearable Accelerometer Nodes