基于最大类可分离性新颜色空间的肤色检测

Skin detection based on maximal class separability in new color space

针对肤色检测,基于Bhattacharyya距离构建了4个新空间模型,并通过计算新空间模型中各特征的Bhattacharyya距离测度,选择其最具鉴别力的分量构成用于肤色分割的联合模型。通过实验比较了肤色、非肤色两类样本在常用的彩色空间和4个新构建空间中的Bhattacharyya距离度量以及肤色正检率,结果表明,基于最大类可分离性判据构建的新彩色空间具有更好的分类性能。在实际彩色图片上的肤色分割实验也证明了提出的新空间模型和联合模型的有效性。

Four new color space models were constructed based on Bhattacharyya distance aiming at skin detection. Meanwhile, according to Bhattacharyya distance measurements of each feature in the new color spaces, some most discernible features were selected to form an assembled model for skin detection. Then, in traditional color spaces and four new spaces, Bhattacharyya distance and positive detection rate were compared on skin and non-skin samples. The experimental results indicate that the new color spaces based on maximal class separability performs better. Eventually, the illustration of skin segmentation in color images also verifies the effectivenesses of both new color space model and assembled model.