OverFeat: Integrated Recognition, Localization and Detection using Convolutional Networks
Preprint
Abstract
We present an integrated framework for using Convolutional Networks for classification, localization and detection. We show how a multiscale and sliding window approach can be efficiently implemented within a ConvNet. We also introduce a novel deep learning approach to localization by learning to predict object boundaries. Bounding boxes are then accumulated rather than suppressed in order to increase detection confidence. We show that different tasks can be learned simultaneously using a single shared network. This integrated framework is the winner of the localization task of the ImageNet Large Scale Visual Recognition Challenge 2013 (ILSVRC2013) and obtained very competitive results for the detection and classifications tasks. In post-competition work, we establish a new state of the art for the detection task. Finally, we release a feature extractor from our best model called OverFeat.Keywords
All Related Versions
This publication has 13 references indexed in Scilit:
- Selective Search for Object RecognitionInternational Journal of Computer Vision, 2013
- Multi-column deep neural networks for image classification2012
- Object Recognition by Sequential Figure-Ground RankingInternational Journal of Computer Vision, 2011
- Traffic sign recognition with multi-scale Convolutional Networks2011
- Constrained parametric min-cuts for automatic object segmentation2010
- Learning long‐range vision for autonomous off‐road drivingJournal of Field Robotics, 2009
- Supervised Learning of Image Restoration with Convolutional Networks2007
- Synergistic Face Detection and Pose Estimation with Energy-Based Models2006
- Gradient-based learning applied to document recognitionProceedings of the IEEE, 1998
- Reading handwritten digits: a ZIP code recognition systemComputer, 1992