首页> 外文期刊>International Journal of Advanced Robotic Systems >A New Profile Shape Matching Stereovision Algorithm for Real-time Human Pose and Hand Gesture Recognition:
【24h】

A New Profile Shape Matching Stereovision Algorithm for Real-time Human Pose and Hand Gesture Recognition:

机译:一种新的轮廓形状匹配立体视觉算法,用于实时人体姿势和手势识别:

获取原文
           

摘要

This paper presents a new profile shape matching stereovision algorithm that is designed to extract 3D information in real time. This algorithm obtains 3D information by matching profile intensity shapes of each corresponding row of the stereo image pair. It detects the corresponding matching patterns of the intensity profile rather than the intensity values of individual pixels or pixels in a small neighbourhood. This approach reduces the effect of the intensity and colour variations caused by lighting differences. As with all real-time vision algorithms, there is always a trade-off between accuracy and processing speed. This algorithm achieves a balance between the two to produce accurate results for real-time applications. To demonstrate its performance, the proposed algorithm is tested for human pose and hand gesture recognition to control a smart phone and an entertainment system.
机译:本文提出了一种新的轮廓形状匹配立体视觉算法,该算法旨在实时提取3D信息。该算法通过匹配立体图像对的每个对应行的轮廓强度形状来获得3D信息。它检测强度分布图的对应匹配模式,而不是单个像素或小邻域中像素的强度值。这种方法减少了由照明差异引起的强度和颜色变化的影响。与所有实时视觉算法一样,在精度和处理速度之间总是要取舍。该算法在两者之间取得了平衡,可以为实时应用生成准确的结果。为了证明其性能,测试了所提出的算法的人体姿势和手势识别,以控制智能手机和娱乐系统。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号