SCNet: Learning Semantic Correspondence

2017/05/11 by Kai Han, Han, Kai, Rafael S. Rezende +11 · 5 citations
Computer Science · #Computer Vision and Pattern Recognition (cs.CV) #FOS: Computer and information sciences #cs.CV

paper · pdf · doi:10.48550/arxiv.1705.04043

ICCV 2017

arxiv created 2017/08/17 · arxiv updated 2017/08/18

Abstract

This paper addresses the problem of establishing semantic correspondences between images depicting different instances of the same object or scene category. Previous approaches focus on either combining a spatial regularizer with hand-crafted features, or learning a correspondence model for appearance only. We propose instead a convolutional neural network architecture, called SCNet, for learning a geometrically plausible model for semantic correspondence. SCNet uses region proposals as matching primitives, and explicitly incorporates geometric consistency in its loss function. It is trained on image pairs obtained from the PASCAL VOC 2007 keypoint dataset, and a comparative evaluation on several standard benchmarks demonstrates that the proposed approach substantially outperforms both recent deep learning architectures and previous methods based on hand-crafted features.

Cited by

• Dense Semantic Matching with VGGT Prior
• Towards Robust Semantic Correspondence: A Benchmark and Insights
• Jamais Vu: Exposing the Generalization Gap in Supervised Semantic Correspondence
• Do It Yourself: Learning Semantic Correspondence from Pseudo-Labels
• Semantic Correspondence: Unified Benchmarking and a Strong Baseline

Related

• SFNet: Learning Object-aware Semantic Correspondence
• Learning Semantic Correspondence Exploiting an Object-level Prior
• Universal Correspondence Network
• Semantic Correspondence: Unified Benchmarking and a Strong Baseline
• AnchorNet: A Weakly Supervised Network to Learn Geometry-sensitive Features For Semantic Matching