Laplacian Pyramid Reconstruction and Refinement for Semantic Segmentation

Abstract

CNN architectures have terrific recognition performance but rely on spatial pooling which makes it difficult to adapt them to tasks that require dense, pixel-accurate labeling. This paper makes two contributions: (1) We demonstrate that while the apparent spatial resolution of convolutional feature maps is low, the high-dimensional feature representation contains significant sub-pixel localization information. (2) We describe a multi-resolution reconstruction architecture based on a Laplacian pyramid that uses skip connections from higher resolution feature maps and multiplicative gating to successively refine segment boundaries reconstructed from lower-resolution maps. This approach yields state-of-the-art semantic segmentation results on the PASCAL VOC and Cityscapes segmentation benchmarks without resorting to more complex random-field inference or instance detection driven architectures.

References in corpus (7)

• Very Deep Convolutional Networks for Large-Scale Image Recognition
• Efficient Inference in Fully Connected CRFs with Gaussian Edge Potentials
• Deep Generative Image Models using a Laplacian Pyramid of Adversarial Networks
• The Cityscapes Dataset for Semantic Urban Scene Understanding
• Learning Deconvolution Network for Semantic Segmentation
• BoxSup: Exploiting Bounding Boxes to Supervise Convolutional Networks for Semantic Segmentation
• Pixel-level Encoding and Depth Layering for Instance-level Semantic Labeling

Cited by in corpus (5)

• DeepLab: Semantic Image Segmentation with Deep Convolutional Nets, Atrous Convolution, and Fully Connected CRFs
• ContextNet: Exploring Context and Detail for Semantic Segmentation in Real-time
• A Survey on Deep Learning Methods for Robot Vision
• Deep Learning For Face Recognition: A Critical Analysis
• Deep Inception-Residual Laplacian Pyramid Networks for Accurate Single Image Super-Resolution