基于手指骨架提取的手形识别方法

来源期刊:中南大学学报(自然科学版)2016年第3期

论文作者:刘富 李温温 高雷

文章页码:777 - 784

关键词:手形识别;手指细化;骨架拟合;手指几何特征

Key words:hand-shape recognition; finger refinement; fitting finger skeleton; finger geometric features

摘    要:针对使用非接触式手形采集仪采集手形时,手掌图像的指根处轮廓会因手掌张开幅度不同而出现不同程度的变形,增加手形精确定位和手形识别的难度这一问题,提出一种不受指根处轮廓变形影响的手形识别算法。该算法首先提取出细化的手指骨架,其次拟合手指骨架并定位手指中线,然后提取与指根部无关的指形特征,最后,计算注册样本与待识别样本的差异特征统计值并进行判别决策。研究结果表明:本文提出的手形识别算法对指根处轮廓变形有着较强的鲁棒性,识别率可达99.62%。

Abstract: Whena non-contact hand shape acquisition instrument is used to collect hand shape, there will be a kind of deformation at different levels in the finger root contour because of the difference in open degree of palm, which increases the difficulty of the accurate positioning and recognition hand-shape. In order to solve this problem, a newmethod for hand-shape recognition was presented, which was not affected by the deformation of the finger root contour. Firstly, the thinning finger skeleton was extracted, and then the finger central axis and the binomial fitting finger skeleton were located. The finger geometric features unrelated to the finger root contour were extracted. Finally, the statisticsof differences characteristic between the registered samples and identifying samples were calculated to recognize the hand shape.The results show that the proposed method is robust to the deformation of the finger root contour, and the recognition rate can reach 99.62%.

有色金属在线官网  |   会议  |   在线投稿  |   购买纸书  |   科技图书馆

中南大学出版社 技术支持 版权声明   电话:0731-88830515 88830516   传真:0731-88710482   Email:administrator@cnnmol.com

互联网出版许可证:(署)网出证(京)字第342号   京ICP备17050991号-6      京公网安备11010802042557号