Improving Deep Image Clustering with Spatial Transformer Layers

Abstract

Image clustering is an important but challenging task in machine learning. As in most image processing areas, the latest improvements came from models based on the deep learning approach. However, classical deep learning methods have problems to deal with spatial image transformations like scale and rotation. In this paper, we propose the use of visual attention techniques to reduce this problem in image clustering methods. We evaluate the combination of a deep image clustering model called Deep Adaptive Clustering (DAC) with the Spatial Transformer Networks (STN). The proposed model is evaluated in the datasets MNIST and FashionMNIST and outperformed the baseline model.

End-to-end deep representation learning for time series clustering: a comparative study

“…Attention mechanisms were first used in Natural Language Processing (NLP) (Bahdanau et al, 2014;Vaswani et al, 2017) then in computer vision (Guan et al, 2018;Jaderberg et al, 2015;Woo et al, 2018). Some methods have been proposed for unsupervised learning, but mainly on specific cases, like graph clustering (Wang et al, 2019) or the use of spatial attention (Souza and Zanchettin, 2019). However, a few works have also been conducted for time series clustering.…”