Timezone: »

Combining Low-Density Separators with CNNs
Yu-Xiong Wang · Martial Hebert

Wed Dec 07 09:00 AM -- 12:30 PM (PST) @ Area 5+6+7+8 #140 #None

This work explores CNNs for the recognition of novel categories from few examples. Inspired by the transferability analysis of CNNs, we introduce an additional unsupervised meta-training stage that exposes multiple top layer units to a large amount of unlabeled real-world images. By encouraging these units to learn diverse sets of low-density separators across the unlabeled data, we capture a more generic, richer description of the visual world, which decouples these units from ties to a specific set of categories. We propose an unsupervised margin maximization that jointly estimates compact high-density regions and infers low-density separators. The low-density separator (LDS) modules can be plugged into any or all of the top layers of a standard CNN architecture. The resulting CNNs, with enhanced generality, significantly improve the performance in scene classification, fine-grained recognition, and action recognition with small training samples.

Author Information

Yu-Xiong Wang (Carnegie Mellon University)
Martial Hebert (Carnegie Mellon University)

More from the Same Authors