改进的Trim-ICP算法在面部整容中的应用  

Application of the Improved Trim-ICP Algorithm in the Facial Surgery

作　　者：朱锋博 郑顺义[1] 王佳婧 ZHU Fengbo;ZHENG Shunyi;WANG Jiajing(School of Remote Sensing and Information Engineering,Wuhan University,Wuhan 430079,China)

机构地区：[1]武汉大学遥感信息工程学院

出　　处：《测绘地理信息》2018年第6期82-85,共4页Journal of Geomatics

基　　金：国家自然科学基金资助项目(41701532);中国博士后科学基金资助项目(2017M612510);中央高校基本科研业务费专项资助项目(2042018gf0003)

摘　　要：在面部整形中,对比整容前后的二维图像可以观察到面部的改变,但是无法进行手术部位在手术前后改变的三维对比展示及定量化描述。使用三维扫描仪器可以获得整容前后人面部的三维点云,利用点云配准算法使点云归化到相同的坐标系下并使未经手术部位的点云重合,从而可以将手术部位凸显出来及进行手术前后面部的变化分析和对比。考虑到整容手术前后,未经手术的部位也可能由于表情变化以及相邻于手术部位的原因而有轻微的改变,在Trim迭代最近点算法的基础上,根据相关点对间的测度将点对进行加权,以使改变越小的部位点云配准得越好,改变越大的部位对比越明显。试验结果表明,该算法能够在准确配准的基础上定量地反映整容前后手术部位三维表面模型的变化。In the facial plastic surgery,the change of the face surgical part can be shown by the comparison of the two - dimension images shot before and after plastic surgery,but it cannot be shown in the three-dimension space and be quanti- tatively described.Three-dimensional scanning instrument can be used to obtain facial point cloud data,when the point cloud data from the part without surgery is aligned by the method of point cloud registration,the change of the surgical part can be highlighted and this change can be analysed and compared.Since the point cloud model of the part without surgery has a slight change due to the expression change and being adjacent to the surgery part,by distributing the weight to the corresponding points this improved Trim-ICP algo- rithm aligns the point cloud of the part which undergoes smallest change better and contrasts the part which under- goes greatest change sharply.The results show that for the face surgical data,the improved Trim-ICP method can get good registration,and the change of the surgical 15art surface mode can be reflected better in quantity.

关 键 词：整容手术 三维点云 改进的Trim迭代最近点算法 

分 类 号：TP751[自动化与计算机技术—检测技术与自动化装置] P208[自动化与计算机技术—控制科学与工程]