Stable Median Centre Clustering for Unsupervised Domain Adaptation Person Re-Identification

Algorithm 1

Stable median centre clustering.

	Input: Source dataset , Target dataset , Maximum iteration N, Style Transfer Model T
	Output: Model
(1)	Transfer the image style of to by T
(2)	Pretrain model on
(3)	for i = 1 to N do
(4)	Extract feature representation from by
(5)	Calculate the pairwise distance between each pair of samples
(6)	Calculate the distance between each pair of clusters according to equation (4), and generate an distance matrix
(7)	Merge clusters and generate pseudolabels by clustering
(8)	Select credible samples with high confidence according to equations (1)–(3)
(9)	Fine-tune model with the credible sample dataset
(10)	end for