To properly detect the face region and improve the accuracy and real-time of face detection and location method in the circumstance of inadequate light of vehicle cab, face detecting method based on skin color clustering is used to segment the color image into skin and non-skin areas. Meanwhile, a new algorithm based on Retinex algorithm is proposed to illumination compensation before the face detection. Simulation experiment has been done with the improved algorithm applied to the newly created database of face image. The detection correctness rate is increased by 4. 7 % when the algorithm is compared with the traditional method. The experimental result prove that this algorithm has good adaptability to the images containing face rotations with different illumination intensity, and it has excellent practical value.%为了正确检测人脸区域、提高驾驶室内光照不足情况下人脸检测与定位方法的准确性和实时性,采用了肤色聚类的人脸检测方法,利用肤色聚类性将彩色图像分割成皮肤区和非皮肤区.同时,提出一种基于多尺度Retinex算法的改进算法,其能够在人脸检测之前对图像进行光照补偿处理.将改进后的算法应用到新建立的人脸图像库中进行仿真实验,并与传统的肤色聚类人脸检测方法进行对比,其正确率提高了4.7％.实验结果表明:改进后的肤色聚类人脸检测算法可实现对不同光照变化和旋转角度的人脸进行检测,且具有很强的实用价值.
展开▼
