Context-aware Padding for Semantic Segmentation
arXiv:2109.07854
Abstract
Zero padding is widely used in convolutional neural networks to prevent the size of feature maps diminishing too fast. However, it has been claimed to disturb the statistics at the border. As an alternative, we propose a context-aware (CA) padding approach to extend the image. We reformulate the padding problem as an image extrapolation problem and illustrate the effects on the semantic segmentation task. Using context-aware padding, the ResNet-based segmentation model achieves higher mean Intersection-Over-Union than the traditional zero padding on the Cityscapes and the dataset of DeepGlobe satellite imaging challenge. Furthermore, our padding does not bring noticeable overhead during training and testing.
References in corpus (10)
- Conditional Image Generation with PixelCNN Decoders
- Pixel Recurrent Neural Networks
- DeepGlobe 2018: A Challenge to Parse the Earth through Satellite Images
- ParseNet: Looking Wider to See Better
- A Review on Deep Learning Techniques Applied to Semantic Segmentation
- OCNet: Object Context Network for Scene Parsing
- ICNet for Real-Time Semantic Segmentation on High-Resolution Images
- Partial Convolution based Padding
- Boundless: Generative Adversarial Networks for Image Extension
- Omnidirectional CNN for Visual Place Recognition and Navigation