The problem of unsupervised and semi-supervised clustering is extensively studied in machine learning. In order to involve user in image data clustering, we proposed in [1] a new approach for interactive semi-supervised clustering that translates user feedback (expressed at the level of individual images) into pairwise constraints between groups of images, these groups being formed thanks to the underlying hierarchical clustering solution and user feedback. Recently, the need for appropriate measures of distance or similarity between data led to the emergence of distance metric learning approaches. In this paper¹, we propose a method incorporating metric learning in the existing system to improve performance and reduce the computational time. Our preliminary experiments performed on the Wang dataset show that metric learning methods improve the performances and computational time of the existing system.