Recognition by association via learning per-exemplar distances

We pose the recognition problem as data association. In this setting, a novel object is explained solely in terms of a small set of exemplar objects to which it is visually similar. Inspired by the work of Frome et al., we learn separate distance functions for each exemplar; however, our distances are interpretable on an absolute scale and can be thresholded to detect the presence of an object. Our exemplars are represented as image regions and the learned distances capture the relative importance of shape, color, texture, and position features for that region. We use the distance functions to detect and segment objects in novel images by associating the bottom-up segments obtained from multiple image segmentations with the exemplar regions. We evaluate the detection and segmentation performance of our algorithm on real-world outdoor scenes from the LabelMe (B. Russel, et al., 2007) dataset and also show some promising qualitative image parsing results.

• Publication year

  2008

• Venue

  2008 IEEE Conference on Computer Vision and Pattern Recognition

• Publication date

  2008-06-23

• Fields of study

  Mathematics, Computer Science

• Identifiers
  DOI 10.1109/CVPR.2008.4587462
• External record

  Open on Semantic Scholar

• Source metadata

  Semantic Scholar

• No claims are published for this paper.

• No concepts are published for this paper.

• Object and scene recognition in tiny images

  2010cited by this paper
• LabelMe: A Database and Web-Based Tool for Image Annotation

  2008influential reference
• Scene Completion Using Millions of Photographs

  2007cited by this paper
• Object Recognition by Scene Alignment

  2007cited by this paper
• Learning Globally-Consistent Local Distance Functions for Shape-Based Image Retrieval and Classification

  2007influential reference
• Tiny images

  2007cited by this paper
• Objects in Context

  2007cited by this paper
• Improving Spatial Support for Objects via Multiple Segmentations

  2007influential reference
• An Exemplar Model for Learning Object Classes

  2007cited by this paper
• The proactive brain: using analogies and associations to generate predictions.

  2007cited by this paper
• The Layout Consistent Random Field for Recognizing and Segmenting Partially Occluded Objects

  2006cited by this paper
• SVM-KNN: Discriminative Nearest Neighbor Classification for Visual Category Recognition

  2006cited by this paper
• Using Multiple Segmentations to Discover Objects and their Extent in Image Collections

  2006cited by this paper
• TextonBoost: Joint Appearance, Shape and Context Modeling for Multi-class Object Recognition and Segmentation

  2006cited by this paper
• Geometric context from a single image

  2005cited by this paper
• Histograms of oriented gradients for human detection

  2005cited by this paper
• Combined Object Categorization and Segmentation With an Implicit Shape Model

  2004cited by this paper
• Object Detection Using the Statistics of Parts

  2004cited by this paper
• Learning Generative Visual Models from Few Training Examples: An Incremental Bayesian Approach Tested on 101 Object Categories

  2004cited by this paper
• Contextual Priming for Object Detection

  2003cited by this paper
• Mean Shift: A Robust Approach Toward Feature Space Analysis

  2002cited by this paper
• Robust Real-Time Face Detection

  2001cited by this paper
• On seeing stuff: the perception of materials by humans and machines

  2001cited by this paper
• Robust Real-time Object Detection

  2001cited by this paper
• Normalized cuts and image segmentation

  1997cited by this paper

• Prototype-Based Semantic Segmentation

  2024cites this paper
• An Image-based Typology for Visualization

  2024cites this paper
• A Neurosymbolic Framework for Bias Correction in Convolutional Neural Networks

  2024cites this paper
• Rethinking Semantic Segmentation: A Prototype View

  2022cites this paper
• Not As Easy As You Think - Experiences and Lessons Learnt from Trying to Create a Bottom-Up Visualization Image Typology

  2022cites this paper
• Region Similarity Representation Learning

  2021cites this paper
• Extended set of superpixel features

  2021cites this paper
• Exemplar-Based Early Event Prediction in Video

  2021cites this paper
• Contextual Image Parsing via Panoptic Segment Sorting

  2021cites this paper
• Large-Scale Unsupervised Semantic Segmentation

  2021cites this paper
• Rethinking preventing class-collapsing in metric learning with margin-based losses

  2020cites this paper
• Reducing Class Collapse in Metric Learning with Easy Positive Sampling

  2020cites this paper
• What Should Not Be Contrastive in Contrastive Learning

  2020cites this paper
• C ONTEXTUAL I MAGE P ARSING VIA P ANOPTIC S EGMENT S ORTING

  2020cites this paper
• Perceptually motivated algorithm for grouping objects in complex scenes

  2019cites this paper
• Machine Perception—Machine Perception MU

  2019cites this paper
• Cooperative Embeddings for Instance, Attribute and Category Retrieval

  2019cites this paper
• SegSort: Segmentation by Discriminative Sorting of Segments

  2019cites this paper
• Invariant Feature Mappings for Generalizing Affordance Understanding Using Regularized Metric Learning

  2019cites this paper
• DeeperLab: Single-Shot Image Parser

  2019cites this paper
• Prototypicality Effects in Global Semantic Description of Objects

  2018cites this paper
• Improving Generalization via Scalable Neighborhood Component Analysis

  2018cites this paper
• BAG-OF-SHAPES DESCRIPTOR USING SHAPE ASSOCIATION BASED ON FREEMAN CHAIN CODE

  2017cites this paper
• Object Detection and Recognition for Assistive Robots: Experimentation and Implementation

  2017cites this paper
• SuperDepthTransfer: Depth Extraction from Image Using Instance-Based Learning with Superpixels

  2017cites this paper
• Depth estimation from multi-scale SLIC superpixels using non-parametric learning

  2017cites this paper
• Object Detection Based on Improved Exemplar SVMs Using a Generic Object Measure

  2017cites this paper
• Holistic image understanding with deep learning and dense random fields

  2016cites this paper
• Semantic segmentation of images exploiting DCT based features and random forest

  2016cites this paper
• Visual Genome: Connecting Language and Vision Using Crowdsourced Dense Image Annotations

  2016cites this paper
• Correlated and Individual Multi-Modal Deep Learning for RGB-D Object Recognition

  2016cites this paper
• Supervision Beyond Manual Annotations for Learning Visual Representations

  2016cites this paper
• Outfit of Exemplar-SVMs for Object Detection and Beyond

  2016cites this paper
• Visipedia circa 2015

  2016cites this paper
• Recognition and detection of objects using visual and textual cues

  2016cites this paper
• Monocular Depth Estimation of Outdoor Scenes Using RGB-D Datasets

  2016cites this paper
• Extracting Visual Knowledge from the Internet: Making Sense of Image Data

  2016cites this paper
• Improving object classification robustness in RGB-D using adaptive SVMs

  2016cites this paper
• Multi-class object segmentation based on jointly integrating segment-level and image-level object priors

  2016cites this paper
• The Discovery of Perceptual Structure from Visual Co-occurrences in Space and Time

  2016cites this paper
• Improving Segmentation Boundaries with Nonparametric Image Parsing

  2015cites this paper
• Unsupervised Texture Segmentation of Natural Scene Images Using Region-based Markov Random Field

  2015cites this paper
• DR AF T Supervision Beyond Human Annotations for Learning Visual Representations

  2015cites this paper
• Towards 4 D Crop Analysis in Precision Agriculture : Estimating Plant Height and Crown Radius over Time via Expectation-Maximization

  2015cites this paper
• Structured prediction for urban scene semantic segmentation with geographic context

  2015cites this paper
• Robust tracking with per-exemplar support vector machine

  2015cites this paper
• Data-Driven Objectness

  2015cites this paper
• Learning to propose objects

  2015cites this paper
• Weakly Supervised Graph Based Semantic Segmentation by Learning Communities of Image-Parts

  2015cites this paper
• Free-Form Region Description with Second-Order Pooling

  2015cites this paper
• Query Adaptive Similarity Measure for RGB-D Object Recognition

  2015cites this paper
• Predicting Images using Convolutional Networks: Visual Scene Understanding with Pixel Maps

  2015cites this paper
• Segmentation and recognition of roadway assets from car-mounted camera video streams using a scalable non-parametric image parsing method

  2015cites this paper
• Object detection based on ensemble of exemplars

  2015cites this paper
• Toward Scene Recognition by Discovering Semantic Structures and Parts

  2015influential citation
• Fast Image Classification for Monument Recognition

  2015cites this paper
• Efficient aspect object models using pre-trained convolutional neural networks

  2015cites this paper
• Data Driven Hierarchical Serial Scene Classification Framework

  2014cites this paper
• HVS-Inspired, Parallel, and Hierarchical Traffic Scene Classification Framework

  2014cites this paper
• Classemes: A Compact Image Descriptor for Efficient Novel-Class Recognition and Search

  2014cites this paper
• Detect2Rank: Combining Object Detectors Using Learning to Rank

  2014cites this paper
• Visual Scene Understanding through Semantic Segmentation

  2014cites this paper
• Research on HVS-Inspired, Parallel, and Hierarchical Scene Classification Framework

  2014cites this paper
• Composite Statistical Learning and Inference for Semantic Segmentation

  2014cites this paper
• VIDEO OBJECT TRACKING BASED ON AUTOMATIC BACKGROUND SEGMENTATION : A SURVEY

  2014cites this paper
• Object Recognition and Semantic Scene Labeling for RGB-D Data

  2014influential citation
• Online Data Clustering Using Variational Learning of a Hierarchical Dirichlet Process Mixture of Dirichlet Distributions

  2014cites this paper
• The SUN Attribute Database: Beyond Categories for Deeper Scene Understanding

  2014cites this paper
• Contextual object category recognition for RGB-D scene labeling

  2014cites this paper
• Learning Everything about Anything: Webly-Supervised Visual Concept Learning

  2014cites this paper
• Fine-Grained Visual Comparisons with Local Learning

  2014cites this paper
• Superpixel Graph Label Transfer with Learned Distance Metric

  2014cites this paper
• Weakly-supervised region annotation for understanding scene images

  2014cites this paper
• Compound Exemplar Based Object Detection by Incremental Random Forest

  2014cites this paper
• E2LSH based multiple kernel approach for object detection

  2014cites this paper
• Dynamic Environment Sensing intelligent vehicle (IV) dynamic environment sensing Using an Intelligent Vehicle intelligent vehicle (IV)

  2013cites this paper
• Mid-level Visual Element Discovery as Discriminative Mode Seeking

  2013cites this paper
• Scene Collaging: Analysis and Synthesis of Natural Images with Semantic Layers

  2013cites this paper
• Beyond visual features: A weak semantic image representation using exemplar classifiers for classification

  2013cites this paper
• Scene image retrieval via re-ranking semantic and packed dense interestpoints

  2013cites this paper
• Towards Open-Universe Image Parsing with Broad Coverage

  2013cites this paper
• Expanding the breadth and detail of object recognition

  2013cites this paper
• RGB-D Object Recognition: Features, Algorithms, and a Large Scale Benchmark

  2013cites this paper
• Runtime Model Recommendation for Exemplar-based Object Detection

  2013cites this paper
• Object class detection: A survey

  2013cites this paper
• Mixtures of Large Margin Nearest Neighbor Classifiers

  2013cites this paper
• Learning Class-to-Image Distance with Object Matchings

  2013influential citation
• Scene Parsing Using Scene Attributes As Global Features

  2013cites this paper
• ImageSpirit

  2013cites this paper
• Nonparametric Scene Parsing with Adaptive Feature Relevance and Semantic Context

  2013cites this paper
• An Object Co-occurrence Assisted Hierarchical Model for Scene Understanding

  2012cites this paper
• A Geometric take on Metric Learning

  2012cites this paper
• Parallel Application Library for Object Recognition

  2012cites this paper
• Recognizing object manipulation activities using depth and visual cues

  2012cites this paper
• Semantic segmentation using regions and parts

  2012cites this paper
• Superpixel level object recognition under local learning framework

  2012cites this paper
• Group-Sensitive Multiple Kernel Learning for Object Recognition

  2012cites this paper
• Bottom-up object segmentation for visual recognition

  2012cites this paper
• Per-Exemplar Fusion Learning for Video Retrieval and Recounting

  2012influential citation
• Measuring Image Distances via Embedding in a Semantic Manifold

  2012cites this paper