Skip to Main Content
This paper presented a special design and implementation of human detection based on SVM (support vector machine) and this method is used in intelligent video surveillance system. In order to simplify the design of the SVM classifier and improve efficiency of machine learning, both a grid vector representation and a center radiating vector representation are proposed to abstract features of the object. The sample data is obtained through processing and analysis including human and no-human which forms the training input to SVM. Finally, we used the trained recognizer to identify whether there is somebody broken into the object region. If there is, the automatic warning device gives the alarm, which guarantees a real-time surveillance.