This paper is inspired from various boundary determination techniques which are used for segregating colours between background, skin and lips. Basic concept for this technique is based on colour segmentation with CIELAB colourspaceudutilized for justifiable reasons. Using LAB colour-space, lips colours were compiled into a colour-map and processedudaccordingly to our proposed algorithm of adaptive circularudenclosure. Algorithm output was determined as a series ofudcoordinates symbolizing boundary values surrounding colourmap. Separation of colours is based on these boundaries by creating a freeform polygon that defines if colour-value either belongs within colour-boundary polygon or not. This technique is famously known as the point in-polygon technique. Proposed technique evaluation uses XM2VTS database based on false positive and false-negative to compute segmentation error. Simulation shows proposed algorithm yields segmented error of 5.55% with accuracy of 94.45%.ud
展开▼
机译：本文的灵感来自各种边界确定技术，这些技术用于在背景，皮肤和嘴唇之间隔离颜色。该技术的基本概念基于具有合理理由的CIELAB色彩空间 udutiled的色彩分割。使用LAB色彩空间，将嘴唇的颜色编译成一个颜色图，并根据我们提出的自适应圆形不封闭算法对它们进行处理。确定算法输出为一系列 udcoordinates，它们表示颜色图周围的边界值。通过创建可定义颜色值是否属于颜色边界多边形内的自由多边形来基于这些边界进行颜色分离。该技术被称为点多边形技术。提议的技术评估使用基于误报和误报的XM2VTS数据库来计算分割错误。仿真表明，该算法产生的分段误差为5.55％，准确度为94.45％。 ud
展开▼
