液晶与显示, 2015, 30 (4): 722, 网络出版: 2016-02-02   

基于混合颜色空间与码本背景减除的虚拟触控优化技术

Touch technology based on mixed color space and codebook background difference
作者单位
中山大学 物理科学与工程技术学院,广东 广州510275
摘要
为了更好实现3D虚拟键盘对人手型的分割,建立了一套基于颜色空间和背景减除的手型分割方法。对原有的单颜色空间处理方法进行了改良与创新,提高运算速度和鲁棒性。介绍了原有颜色空间算法的原理和步骤,并给出该方法的效果图和运算速度。根据颜色空间处理方法的原理分析程序中可以提高速度的部分,提出一些颜色空间算法无法解决的问题。在原有HSV颜色空间的处理中加入新颜色空间YCrCb,结合两者进行速度方面的提升,最后,在颜色空间处理之外,加入码本法背景减除,解决干扰色、手饰品干扰等颜色空间无法处理的问题。实验结果表明: 使用新颜色空间算法,运算速度提升,节省28.27%的时间; 加入码本背景减除后的指尖定位良好,平均误差小于2%。达到了提高3D虚拟键盘实时性和鲁棒性的要求。
Abstract
In order to improve the skin detection in 3D virtual keyboard, this paper establishes a new skin detection method based on the color space and the background difference.Both the robustness and the speed of the skin detection are promoted after the replacement of the origin method by the new method.Firstly, this paper presents the theory and the detailed steps, and gives the results of the previous method.Then, some problems are presented and the paper gives much advice about improving the speed by analyzing the theory.Combining the HSV color space with the YCrCb color space helps to improve the operating speed.Also, the paper makes the background difference as an adjunction to the color space, and solve the problems, such as the negative effects of the interference colors and the jewelries, that cannot be overcome just by the color space.The experience shows that, after applying the new method, the operation time is shortened by 28.27% and the average error of finger recognition is lower than 2%.All of these indicates that the new method meets the requirement of promoting the robustness and the speed of the skin detection.

林孜铠, 戴思璋, 梁浩文, 林岱昆, 朱洁霞, 苏剑邦, 陈海域, 王嘉辉, 周建英. 基于混合颜色空间与码本背景减除的虚拟触控优化技术[J]. 液晶与显示, 2015, 30(4): 722. LIN Zi-kai, DAI Si-zhang, LIANG Hao-wen, LIN Dai-kun, ZHU Jie-xia, SU Jian-bang, CHEN Hai-yu, WANG Jia-hui, ZHOU Jian-ying. Touch technology based on mixed color space and codebook background difference[J]. Chinese Journal of Liquid Crystals and Displays, 2015, 30(4): 722.

本文已被 1 篇论文引用
被引统计数据来源于中国光学期刊网
引用该论文: TXT   |   EndNote

相关论文

加载中...

关于本站 Cookie 的使用提示

中国光学期刊网使用基于 cookie 的技术来更好地为您提供各项服务,点击此处了解我们的隐私策略。 如您需继续使用本网站,请您授权我们使用本地 cookie 来保存部分信息。
全站搜索
您最值得信赖的光电行业旗舰网络服务平台!