A novel evaluation method of Chinese female lower body shapes based on machine learning

Abstract

Purpose

The purpose of this paper is to define the evaluation criteria for Chinese female lower body shapes and simplify the evaluation process of shapewear, including girdles, shaping pants, etc.

Design/methodology/approach

The study utilized a machine learning algorithm based on support vector regression and optimized by a genetic algorithm to construct an evaluation model for the contour beauty of Chinese female lower body shapes. A total of 64 virtual 3D models of women were measured. These models were rated by 42 raters using the Likert 9 psychological scale and data was obtained from 310 female samples.

Findings

Eight factors were selected and used to create a regression prediction model. The model achieved an accuracy of 84.7% for the training samples and 86.6% for the test samples.

Originality/value

The model can be utilized to assess the aesthetic appeal of the female lower body and to evaluate the shaping impact of shapewear. The research and evaluation of shapewear for the female lower body are of great significance, particularly in enhancing production efficiency.