Convolutional Oriented Boundaries

Abstract

We present Convolutional Oriented Boundaries (COB), which produces multiscale oriented contours and region hierarchies starting from generic image classification Convolutional Neural Networks (CNNs). COB is computationally efficient, because it requires a single CNN forward pass for contour detection and it uses a novel sparse boundary representation for hierarchical segmentation; it gives a significant leap in performance over the state-of-the-art, and it generalizes very well to unseen categories and datasets. Particularly, we show that learning to estimate not only contour strength but also orientation provides more accurate results. We perform extensive experiments on BSDS, PASCAL Context, PASCAL Segmentation, and MS-COCO, showing that COB provides state-of-the-art contours, region hierarchies, and object proposals in all datasets.

ECCV 2016 Camera Ready

References in corpus (4)

• Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks
• What makes for effective detection proposals?
• Multiscale Combinatorial Grouping for Image Segmentation and Object Proposal Generation
• Deep Retinal Image Understanding

Cited by in corpus (21)

• Richer Convolutional Features for Edge Detection
• Learning Video Object Segmentation from Static Images
• Lightweight Pixel Difference Networks for Efficient Visual Representation Learning
• On Detection of Faint Edges in Noisy Images
• Learning Regional Attraction for Line Segment Detection
• Bi-Directional Cascade Network for Perceptual Edge Detection
• Oriented Response Networks
• Where are the Masks: Instance Segmentation with Image-level Supervision
• CASENet: Deep Category-Aware Semantic Edge Detection
• Object Counting and Instance Segmentation with Image-level Supervision
• Pixel Difference Networks for Efficient Edge Detection
• Instance Segmentation with Point Supervision
• Image-based localization using LSTMs for structured feature correlation
• Recent Advances in the Applications of Convolutional Neural Networks to Medical Image Contour Detection
• Learning Crisp Boundaries Using Deep Refinement Network and Adaptive Weighting Loss
• Deep Multicameral Decoding for Localizing Unoccluded Object Instances from a Single RGB Image
• Long-Term Image Boundary Prediction
• Probabilistic Model of Visual Segmentation
• Learning to Parse Wireframes in Images of Man-Made Environments
• SUSiNet: See, Understand and Summarize it
• Harvesting Visual Objects from Internet Images via Deep Learning Based Objectness Assessment