arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 1212.0134cs.CV
收藏

指尖检测:一种自然手的快速方法

Fingertip Detection: A Fast Method with Natural Hand

CEERI/CSIR Pilani(中央电子工程研究所/印度科学工业研究理事会(皮拉尼));Don Bosco University(唐博斯科大学);BITS Pilani(皮拉尼比尔拉理工学院)

J. L. Raheja, Karen Das, Ankit Chaudhary

摘要:

许多基于视觉的应用已经在其应用中使用指尖来跟踪或操纵手势。手势识别是一种将信号传递给机器的自然方式,因为人类大部分时间都用手势表达自己的情感。这里描述了一种新的时间有效的算法用于指尖检测。该方法对手的方向具有不变性,并且在预处理时只从整幅图像中分割出手的一部分,因此进一步的计算将比处理整幅图像快得多。该方法首先利用基于HSV颜色空间的皮肤滤波器生成输入图像的二值轮廓,然后根据手部图像的亮度直方图进行手部裁剪

英文摘要:

Many vision based applications have used fingertips to track or manipulate gestures in their applications. Gesture identification is a natural way to pass the signals to the machine, as the human express its feelings most of the time with hand expressions. Here a novel time efficient algorithm has been described for fingertip detection. This method is invariant to hand direction and in preprocessing it cuts only hand part from the full image, hence further computation would be much faster than processing full image. Binary silhouette of the input image is generated using HSV color space based skin filter and hand cropping done based on intensity histogram of the hand image

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑