【摘要】 脸型分类是人脸分析领域中的重要任务，对于人脸识别和美学评估等应用具有重要意义。针对传统脸型分类算法准确率低、分类速度慢的问题，提出一种新的基于Dlib和K-means算法的人脸脸型分类算法。首先创建级联分类器和面部特征点预测器进行面部检测和关键点定位，使用高斯滤波器对灰度图像进行平滑处理，使用K-means算法对额头区域进行聚类，通过测量额头线、颧骨线、下颚线、脸长线的长度和脸颊角度根据算法确定面部形状。根据实验集测试结果，相对于现有算法而言，此算法在脸型分类的准确性显著提升。更多还原

【Abstract】 Face shape classification is an important task in the field of facial analysis, with significant implications for applications such as face recognition and aesthetic assessment. In response to the low accuracy and slow classification speed of traditional face shape classification algorithms, a novel face shape classification algorithm based on Dlib and K-means is proposed. Firstly, a cascade classifier and facial landmark predictor are created for face detection and key point localization. The grayscale image is smoothed using a Gaussian filter, and the forehead region is clustered using the K-means algorithm. Facial shape is determined based on the lengths and angles of the forehead line, cheekbone line, jawline, and face length line measured by the algorithm. Experimental results on the test dataset demonstrate that this algorithm significantly improves the accuracy of face shape classification compared to existing algorithms.更多还原