【摘要】 提出基于小波轮廓描述符和支持向量机的视频对象识别方法,主要根据形状的不同来区分不同的视频运动对象。该方法的识别效率高,可以准确区分人、动物、车辆等不同对象,还可以区分人的站立、卧倒、坐下等不同的姿态,识别性能比较稳定,识别速度快,能够满足实时视频处理的需要,具有良好的理论价值和应用前景。更多还原

【Abstract】 Presents a Video object recognition method based on wavelet contour descriptors and support vector machine,according to the different shape to distinguish between different video motion object.The method for identification of high efficiency,can accurately distinguish people,animals, vehicles and other objects,can also differentiate between people stand,lie down,sit down, different attitude,recognition performance is relatively stable,the recognition speed is fast,able to meet the need of real-time video processing,and has a good theory value and application prospect.更多还原