litlas
ログインアカウント作成

Proposal Flow

著者: Bumsub Ham, Minsu Cho, Cordelia Schmid, Jean Ponce - IEEE/CVF Conference on Computer Vision and Pattern Recognition, CVPR 2016 被引用: 101

Finding image correspondences remains a challenging problem in the presence of intra-class variations and large changes in scene layout. Semantic flow methods are designed to handle images depicting different instances of the same object or scene category. We introduce a novel approach to semantic flow, dubbed proposal flow, that establishes reliable correspondences using object proposals. Unlike prevailing semantic flow approaches that operate on pixels or regularly sampled local regions, proposal flow benefits from the characteristics of modern object proposals, that exhibit high repeatability at multiple scales, and can take advantage of both local and geometric consistency constraints among proposals. We also show that proposal flow can effectively be transformed into a conventional dense flow field. We introduce a new dataset that can be used to evaluate both general semantic flow techniques and region-based approaches such as proposal flow. We use this benchmark to compare different matching algorithms, object proposals, and region features within proposal flow, to the state of the art in semantic flow. This comparison, along with experiments on standard datasets, demonstrates that proposal flow significantly outperforms existing semantic flow methods in various settings.

On SIFTs and theirscalesOn SIFTs and their scalesDeformable SpatialPyramid Matching for…Deformable Spatial Pyramid Matching for Fast Dense CorrespondencesDAISY Filter Flow: AGeneralized Discrete…DAISY Filter Flow: A Generalized Discrete Approach to Dense CorrespondencesScale-Space SIFT flowScale-Space SIFT flowDo Convnets LearnCorrespondence?Do Convnets Learn Correspondence?Spatial Pyramid Poolingin Deep Convolutional…Spatial Pyramid Pooling in Deep Convolutional Networks for Visual RecognitionDense SemanticCorrespondence Where…Dense Semantic Correspondence Where Every Pixel is a ClassifierGeneralized DeformableSpatial Pyramid…Generalized Deformable Spatial Pyramid: Geometry-preserving dense correspondence estimationFlowWeb: Joint image setalignment by weaving…FlowWeb: Joint image set alignment by weaving consistent, pixel-wise correspondencesDense CorrespondencesAcross Scenes and ScalesDense Correspondences Across Scenes and ScalesUnsupervised ObjectDiscovery and…Unsupervised Object Discovery and Localization in the Wild: Part-based Matching with Bottom-up Region ProposalsFast R-CNNFast R-CNNSCNet: Learning SemanticCorrespondenceSCNet: Learning Semantic CorrespondenceDeep Semantic FeatureMatchingDeep Semantic Feature MatchingObject-Aware DenseSemantic CorrespondenceObject-Aware Dense Semantic CorrespondenceConvolutional NeuralNetwork Architecture fo…Convolutional Neural Network Architecture for Geometric MatchingDCTM:Discrete-Continuous…DCTM: Discrete-Continuous Transformation Matching for Semantic FlowFCSS: FullyConvolutional…FCSS: Fully Convolutional Self-Similarity for Dense Semantic CorrespondenceFCSS: FullyConvolutional…FCSS: Fully Convolutional Self-Similarity for Dense Semantic CorrespondenceDiscrete-ContinuousTransformation Matching…Discrete-Continuous Transformation Matching for Dense Semantic CorrespondenceHierarchical semanticimage matching using CN…Hierarchical semantic image matching using CNN feature pyramidLearning SemanticCorrespondence…Learning Semantic Correspondence Exploiting an Object-Level PriorProbabilistic ModelDistillation for…Probabilistic Model Distillation for Semantic CorrespondenceConvolutional HoughMatching Networks for…Convolutional Hough Matching Networks for Robust and Efficient Visual CorrespondenceProposal FlowProposal Flow過去の参考文献中心の論文この論文を引用する論文古い新しい

ノードをクリックするとフォーカスを固定、空白をクリックすると本論文に戻ります。ホバーで一時的にプレビューできます。各ノードのページはタイトルから開けます。