Discriminatively Boosted Image Clustering with Fully Convolutional Auto-Encoders

23 Mar 2017  ·  Fengfu Li, Hong Qiao, Bo Zhang, Xuanyang Xi ·

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.

PDF Abstract

Results from the Paper

Task Dataset Model Metric Name Metric Value Global Rank Result Benchmark
Image Clustering coil-100 DBC NMI 0.905 # 8
Accuracy 0.775 # 5
Image Clustering Coil-20 DBC NMI 0.895 # 4
Accuracy 0.793 # 3
Image Clustering MNIST-full DBC NMI 0.937 # 10
Accuracy 0.976 # 8
Image Clustering USPS DBC NMI 0.724 # 16
Accuracy 0.743 # 13


No methods listed for this paper. Add relevant methods here