@article{xiangfang_research_2026,
	title = {Research on colour matching recommendation for clothing users based on the {DBSCAN} clustering algorithm},
	issn = {1222-5347},
	url = {https://revistaindustriatextila.ro/images/2026/2/008%20REN%20XIANGFANG_INDUSTRIA%20TEXTILA%20no.2_2026.pdf},
	doi = {10.35530/IT.077.02.202510},
	abstract = {Colour plays an important role in promoting the sales relationship between clothing brands and users. This article is
based on publicly available e-commerce image data from Tmall and the brand’s official website. Taking the Taiping Bird
brand as an example, a large amount of image data is preprocessed using the Structural Similarity SSIM index, which
is used to evaluate image similarity. The image data is vectorised using a non-linear conversion formula from RGB
format to HSV format, and K-means clustering analysis is performed to obtain the main colours of the image. Finally, the
BDSCAN clustering algorithm is used to cluster the large amount of colour data, continuously modifying the
neighbourhood threshold eps and minimum sample threshold minPts parameters until a sufficient number of clusters
are obtained and a relatively low noise rate is achieved. Finally, for each category, a greedy algorithm is used to solve
for the minimum subset of samples that can represent each cluster. The results are presented in the form of web pages,
divided into two-page entrances for users and brand owners. It provides competitive brand colour recommendations and
real-time sales to brand owners, and personalised colour preference recommendations and direct links to recommended
products to users.},
	urldate = {2026-05-17},
	journal = {Industria Textila},
	author = {Xiangfang, Ren and Ru, Fan and Lei, Shen},
	month = may,
	year = {2026},
	pages = {259},
}