Unsupervised Learning of Discriminative Attributes and Visual Representations

Abstract

Attributes offer useful mid-level features to interpret visual data. While most attribute learning methods are supervised by costly human-generated labels, we introduce a simple yet powerful unsupervised approach to learn and predict visual attributes directly from data. Given a large unlabeled image collection as input, we train deep Convolutional Neural Networks (CNNs) to output a set of discriminative, binary attributes often with semantic meanings. Specifically, we first train a CNN coupled with unsupervised discriminative clustering, and then use the cluster membership as a soft supervision to discover shared attributes from the clusters while maximizing their separability. The learned attributes are shown to be capable of encoding rich imagery properties from both natural images and contour patches. The visual representations learned in this way are also transferrable to other tasks such as object detection. We show other convincing results on the related tasks of image retrieval and classification, and contour detection.

Method and applications

Pretraining from scratch with modified unsupervised discriminative clustering + finetuning with weakly-supervised hashing

Enable diverse vision applications:

• Unsupervised pre-training for object detection
• Unsupervised learning for image retrieval
• Unsupervised learning for image classification
• Unsupervised learning for contour detection

Citation

@inproceedings{huang2016unsupervised,
 author = {Chen Huang and Chen Change Loy and Xiaoou Tang},
 title = {Unsupervised Learning of Discriminative Attributes and Visual Representations},
 booktitle = {IEEE Conference on Computer Vision and Pattern Recognition (CVPR)},
 year = {2016},}