基于二维照片的青年男性颈肩部形态分类与识别

为探究青年男性颈肩部形态分类并实现基于照片的自动识别,首先通过三维人体扫描仪获取180名男大学生颈肩部的点云数据,测量了22个与男性颈肩部形态相关的特征参数;然后根据变异系数分析选取前倾角、背入角、肩斜角、颈肩宽比、颈横矢径比作为聚类分析变量,对颈肩部形态进行分类并总结判别规则;最后结合人体二维照片提取体型分类所需参数,构建了颈肩部形态自动识别系统。结果显示:青年男性颈肩部形态可分为落肩圆颈体、前倾圆颈体、宽颈直体3类,构建的形态自动识别系统的判别准确率达到93. 33%,说明本文方法可行且有效,可满足消费者个性化定制的需求。

Classification and recognition of young males' neck-shoulder shape based on 2-D photos

In order to explore the classification of the neck and shoulder shape of young men and to facilitate the automatic recognition based on 2-D photos, this study obtained the point cloud data of 180 male college students using a three-dimensional body scanner, and measured 22 characteristic parameters relating to the male neck and shoulder shape. According to the analysis of the coefficient of variation, the forward angle, back angle, shoulder oblique angle, neck-to-shoulder width ratio, and neck transverse sagittal diameter ratio were selected as cluster analysis variables to classify the neck and shoulder shapes for establishing the discriminant rules. Combining the 2-D photos of the human body to extract the parameters required for body type classification, an automatic recognition system for the shape of the neck-shoulders was constructed. The results show that the neck-shoulder shape of young men can be divided into three types, namely round-neck-drop-shoulder, forward-round-neck and wide-neck-straight-body. The discrimination accuracy rate of the constructed form automatic recognition system reached 93.33%, indicating that this method is effective and feasible, and can meet the needs of consumers' personalized customization.