Visualizing and Understanding Convolutional Networks
Matthew D. Zeiler, Rob Fergus · arXiv (Cornell University) · 2013
Large Convolutional Network models have recently demonstrated impressive classification performance on the ImageNet benchmark. However there is no clear understanding of why they perform so well, or how they might be improved. In this paper we address both issues. We introduce a novel visualization technique that gives insight into the function of intermediate feature layers and the operation of the classifier. We also perform an ablation study to discover the performance contribution from different model layers. This enables us to find model architectures that outperform Krizhevsky \etal on the ImageNet classification benchmark. We show our ImageNet model generalizes well to other datasets: when the softmax classifier is retrained, it convincingly beats the current state-of-the-art results on Caltech-101 and Caltech-256 datasets.
Cited by 394 works. doi:10.48550/arxiv.1311.2901 Open-access PDF
Authors
References
- Backpropagation Applied to Handwritten Zip Code Recognition
- A Fast Learning Algorithm for Deep Belief Nets
- Improving neural networks by preventing co-adaptation of feature detectors
- Visualizing and Understanding Convolutional Networks
- What is the best multi-stage architecture for object recognition?
- Greedy Layer-Wise Training of Deep Networks
- Extracting and composing robust features with denoising autoencoders
- One-shot learning of object categories
- Linear spatial pyramid matching using sparse coding for image classification
- Unbiased look at dataset bias