Title: Unsupervised Bayesian method for colour matching of product packaging

Authors: Zhibing Liu; Han Liu

Addresses: Academy of Fine Arts, Huaibei Normal University, Huaibei, 235000, China ' School of Computer Science and Engineering, Northeastern University, Shenyang 110000, China

Abstract: There are some problems in the traditional colour matching methods of product packaging, such as high accuracy of colour matching and long consumption time. This paper proposes a colour matching method of product packaging based on unsupervised Bayes. First, the distance threshold between different colours of product packaging was obtained. Second, the colour features were extracted by sparse representation method, and the colour features were preprocessed by fuzzy c-means clustering. Third, the colour of product outer packaging is sorted according to the proportion to realise the quantification of product outer packaging colour. Then, the colour matching parameters of product packaging are determined by Dirichlet distribution in unsupervised Bayesian. Finally, the optimal colour matching scheme is determined by introducing multiple beta functions. The importance weight of matching scheme is modified by approximate posterior probability to complete the matching. The results show that the minimum error of the proposed method is about 0.5%.

Keywords: product packaging; distance threshold; sparse representation; colour centre value; unsupervised Bayes; collocation.

DOI: 10.1504/IJMTM.2022.128722

International Journal of Manufacturing Technology and Management, 2022 Vol.36 No.5/6, pp.273 - 283

Received: 02 Apr 2021
Accepted: 27 Oct 2021

Published online: 02 Feb 2023 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article