Discriminatively Boosted Image Clustering with Fully Convolutional Auto-Encoders

Abstract

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.

27 pages

References in corpus (5)

• Very Deep Convolutional Networks for Large-Scale Image Recognition
• Learning Deconvolution Network for Semantic Segmentation
• Joint Unsupervised Learning of Deep Representations and Image Clusters
• Deep Deconvolutional Networks for Scene Parsing
• Effective Deterministic Initialization for $k$-Means-Like Methods via Local Density Peaks Searching

Cited by in corpus (6)

• Clustering with Deep Learning: Taxonomy and New Methods
• Identifying Strong Lenses with Unsupervised Machine Learning using Convolutional Autoencoder
• AutoEmbedder: A semi-supervised DNN embedding system for clustering
• Deep Amortized Clustering
• Deep Discriminative Latent Space for Clustering
• A Self-supervised Learning System for Object Detection in Videos Using Random Walks on Graphs