With the development and progress of the times, whether it is the practicality or fashion of clothing, people's requirements for clothing are getting higher and higher. Clothing is an important part of people's daily life. With the improvement of people's overall quality, there are new requirements for the overall style of clothing, such as style, color, fabric comfort, etc. Clothing design has a non-negligible impact on clothing structure and clothing management. In this paper, a collaborative filtering clothing recommendation algorithm based on image visual features is designed. The algorithm uses the matrix decomposition model to obtain the user feature partial favorability matrix and the commodity feature possession matrix through the user-item scoring information. Experiments show that compared with the benchmark algorithm Funk-SVD, the recall, precision, and F1 scores are improved. Therefore, our algorithm can effectively analyze clothing design and clothing structure management, and give better suggestions for people.

随着时代的发展和进步，无论是服装的实用性还是时尚性，人们对服装的要求越来越高。服装是人们日常生活的重要组成部分。随着人们综合素质的提高，对服装的整体风格，如款式、色彩、面料舒适度等有了新的要求，服装设计对服装结构和服装管理有着不可忽视的影响。本文设计了一种基于图像视觉特征的协同过滤服装推荐算法。该算法采用矩阵分解模型，通过用户-物品评分信息得到用户特征偏好感矩阵和商品特征拥有矩阵。实验表明，与基准算法Funk-SVD相比，查全率、查准率和F1分数都有所提高。因此，我们的算法可以有效地分析服装设计和服装结构管理，并为人们提供更好的建议。