Discriminatively Boosted Image Clustering with Fully Convolutional Auto-Encoders
This work addresses image clustering for computer vision applications, presenting an incremental improvement by integrating feature learning and clustering into a single framework with boosted discrimination.
The paper tackles the problem of image clustering by proposing a unified framework that jointly learns image representations and cluster centers using fully convolutional auto-encoders and soft k-means scores, achieving state-of-the-art performance on several vision benchmark datasets.
Traditional image clustering methods take a two-step approach, feature learning and clustering, sequentially. However, recent research results demonstrated that combining the separated phases in a unified framework and training them jointly can achieve a better performance. In this paper, we first introduce fully convolutional auto-encoders for image feature learning and then propose a unified clustering framework to learn image representations and cluster centers jointly based on a fully convolutional auto-encoder and soft $k$-means scores. At initial stages of the learning procedure, the representations extracted from the auto-encoder may not be very discriminative for latter clustering. We address this issue by adopting a boosted discriminative distribution, where high score assignments are highlighted and low score ones are de-emphasized. With the gradually boosted discrimination, clustering assignment scores are discriminated and cluster purities are enlarged. Experiments on several vision benchmark datasets show that our methods can achieve a state-of-the-art performance.