Spatially-Adaptive Filter Units for Compact and Efficient Deep Neural Networks
Preprint
Abstract
Convolutional neural networks excel in a number of computer vision tasks. One of their most crucial architectural elements is the effective receptive field size, that has to be manually set to accommodate a specific task. Standard solutions involve large kernels, down/up-sampling and dilated convolutions. These require testing a variety of dilation and down/up-sampling factors and result in non-compact representations and excessive number of parameters. We address this issue by proposing a new convolution filter composed of displaced aggregation units (DAU). DAUs learn spatial displacements and adapt the receptive field sizes of individual convolution filters to a given problem, thus eliminating the need for hand-crafted modifications. DAUs provide a seamless substitution of convolutional filters in existing state-of-the-art architectures, which we demonstrate on AlexNet, ResNet50, ResNet101, DeepLab and SRN-DeblurNet. The benefits of this design are demonstrated on a variety of computer vision tasks and datasets, such as image classification (ILSVRC 2012), semantic segmentation (PASCAL VOC 2011, Cityscape) and blind image de-blurring (GOPRO). Results show that DAUs efficiently allocate parameters resulting in up to four times more compact networks at similar or better performance.Keywords
All Related Versions
This publication has 19 references indexed in Scilit:
- Structured Receptive Fields in CNNs2016
- Deep Residual Learning for Image Recognition2016
- The Cityscapes Dataset for Semantic Urban Scene Understanding2016
- SOPHIE velocimetry ofKeplertransit candidatesAstronomy & Astrophysics, 2016
- Fully convolutional networks for semantic segmentation2015
- ImageNet Large Scale Visual Recognition ChallengeInternational Journal of Computer Vision, 2015
- Medical Image Computing and Computer-Assisted Intervention -- MICCAI 20152015
- BING: Binarized Normed Gradients for Objectness Estimation at 300fps2014
- Mastering the Discrete Fourier Transform in One, Two or Several Dimensions2013
- Gradient-based learning applied to document recognitionProceedings of the IEEE, 1998