Photo from wikipedia
Digital image processing has been widely used for researches in the fashion industry. This study presented a method to classify women's waist-hip-leg position based on body images. 135 healthy female… Click to show full abstract
Digital image processing has been widely used for researches in the fashion industry. This study presented a method to classify women's waist-hip-leg position based on body images. 135 healthy female students were selected as the experimental subjects, and then photo and 3D body measuring methods were used to obtain 40 shape parameters. Through factor analysis, five factors were extracted for the waist-hip position and eight factors for the leg position. The waist-hip and leg were separately classified into four categories after clustering analysis with optimized factors. The distribution of waist-hip-leg shape was analyzed by combining the waist-hip and leg classification results. The results showed that 12.31% of the subjects had prominent abdomens, flat buttocks, and round and thin legs. The landmarks and parameters were automatically extracted for waist-hip-leg shape identification. The image-based shape analysis approach (ISA) was finally verified with an accuracy rate of over 90% with 30 new subjects.