Do We Need More Training Data?

Datasets for training object recognition systems are steadily increasing in size. This paper investigates the question of whether existing detectors will continue to improve as data grows, or saturate in performance due to limited model complexity and the Bayes risk associated with the feature spaces in which they operate. We focus on the popular paradigm of discriminatively trained templates defined on oriented gradient features. We investigate the performance of mixtures of templates as the number of mixture components and the amount of training data grows. Surprisingly, even with proper treatment of regularization and “outliers”, the performance of classic mixture models appears to saturate quickly (∼10\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${\sim }10$$\end{document} templates and ∼100\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${\sim }100$$\end{document} positive training examples per template). This is not a limitation of the feature space as compositional mixtures that share template parameters via parts and that can synthesize new templates not encountered during training yield significantly better performance. Based on our analysis, we conjecture that the greatest gains in detection performance will continue to derive from improved representations and learning algorithms that can make efficient use of large datasets.

• Publication year

  2015

• Venue

  International Journal of Computer Vision

• Publication date

  2015-03-01

• Fields of study

  Computer Science

• Identifiers
  DOI 10.1007/s11263-015-0812-2arXiv 1503.01508
• External record

  Open on Semantic Scholar

• Source metadata

  Semantic Scholar

• No claims are published for this paper.

• No concepts are published for this paper.

• Face detection, pose estimation, and landmark localization in the wild

  2012influential reference
• Diagnosing Error in Object Detectors

  2012cited by this paper
• Distance Transforms of Sampled Functions

  2012cited by this paper
• ImageNet classification with deep convolutional neural networks

  2012cited by this paper
• How Important Are "Deformable Parts" in the Deformable Parts Model?

  2012cited by this paper
• Ensemble of exemplar-SVMs for object detection and beyond

  2011influential reference
• LIBSVM: A library for support vector machines

  2011influential reference
• Nonparametric Scene Parsing via Label Transfer

  2011cited by this paper
• Unbiased look at dataset bias

  2011cited by this paper
• Finding the weakest link in person detectors

  2011cited by this paper
• Superparsing

  2010cited by this paper
• Object Detection with Discriminatively Trained Part Based Models

  2010influential reference
• What Does Classifying More Than 10, 000 Image Categories Tell Us?

  2010cited by this paper
• The Pascal Visual Object Classes (VOC) Challenge

  2010cited by this paper
• The Unreasonable Effectiveness of Data

  2009cited by this paper
• Fast Approximate Nearest Neighbors with Automatic Algorithm Configuration

  2009cited by this paper
• Poselets: Body part detectors trained using 3D human pose annotations

  2009cited by this paper
• Multiple kernels for object detection

  2009cited by this paper
• Multi-PIE

  2008cited by this paper
• In defense of Nearest-Neighbor based image classification

  2008cited by this paper
• IM2GPS: estimating geographic information from a single image

  2008cited by this paper
• Nearest-Neighbor Methods in Learning and Vision

  2008cited by this paper
• Robust Truncated Hinge Loss Support Vector Machines

  2007cited by this paper
• Image Classification using Random Forests and Ferns

  2007cited by this paper
• Scene Completion Using Millions of Photographs

  2007cited by this paper
• SVM-KNN: Discriminative Nearest Neighbor Classification for Visual Category Recognition

  2006cited by this paper
• Nearest-Neighbor Methods in Learning and Vision: Theory and Practice (Neural Information Processing)

  2006cited by this paper
• Histograms of oriented gradients for human detection

  2005influential reference
• Fast pose estimation with parameter-sensitive hashing

  2003cited by this paper
• Probabilistic Outputs for Support vector Machines and Comparisons to Regularized Likelihood Methods

  1999cited by this paper
• Some PAC-Bayesian Theorems

  1998cited by this paper
• Shape indexing using approximate nearest-neighbour search in high-dimensional spaces

  1997cited by this paper
• Ieee Transactions on Pattern Analysis and Machine Intelligence 1 80 Million Tiny Images: a Large Dataset for Non-parametric Object and Scene Recognition

  year unknowncited by this paper

• Soil class mapping with severely limited sample data: an explicit rule of soil-landscape relationships in complex terrain

  2026cites this paper
• Evaluating large language models for inverse semiconductor design

  2026cites this paper
• Modelling Peatland Productivity by Water Table Depth or Near-Surface Water Contents via the DIMONA Online Platform

  2025cites this paper
• Cloud-based machine learning for scalable classification of software requirements: Insights from the PROMISE dataset

  2025cites this paper
• Large language models for identifying social determinants of health

  2025cites this paper
• Exploring soil temperature extremes: unraveling dynamics with local and spatial machine learning models

  2025cites this paper
• UtilGen: Utility-Centric Generative Data Augmentation with Dual-Level Task Adaptation

  2025cites this paper
• DeepCQ: General-Purpose Deep-Surrogate Framework for Lossy Compression Quality Prediction

  2025cites this paper
• Pomerania Fish: A dataset for fishes across Pomerania freshwater waterbodies in-situ environments

  2025cites this paper
• The third wheel or the game changer? How AI could team up with neurologists in Parkinson's care.

  2025cites this paper
• A fuzzy decision-making network model for offshore wind turbine selection based on simulated annealing algorithm

  2025cites this paper
• Novel Machine Learning Unlocks High Lipid Productivity and Resolves Trade-offs in Algal Biofuel Production

  2025cites this paper
• Improving personalized federated learning to optimize site-specific performance in computer-aided detection/diagnosis

  2025cites this paper
• Dinic’s assembly parts matching optimization based on CACGAN gyroscope performance prediction

  2025cites this paper
• Finding the Sweet Spot: An Empirical Study on Dataset Size, Performance, and Efficiency in Relation Extraction

  2025cites this paper
• Advancing the Prediction and Evaluation of Blast-Induced Ground Vibration Using Deep Ensemble Learning with Uncertainty Assessment

  2025cites this paper
• Error in the Loop: How Human Mistakes Can Improve Algorithmic Learning

  2025cites this paper
• Acoustic Detection of Forest Wood-Boring Insects Under Co-Infestations

  2025cites this paper
• Are Classical Clone Detectors Good Enough for the AI Era?

  2025cites this paper
• Automated detection of sea cucumbers in turbid subtidal marine habitats: An explainable approach

  2025cites this paper
• Flash Flood Risk Classification Using GIS-Based Fractional Order k-Means Clustering Method

  2025cites this paper
• Dataset meta-level and statistical features affect machine learning performance

  2024cites this paper
• Data Deletion for Linear Regression with Noisy SGD

  2024cites this paper
• Generative Active Learning with Variational Autoencoder for Radiology Data Generation in Veterinary Medicine

  2024cites this paper
• Toward Robust Canine Cardiac Diagnosis: Deep Prototype Alignment Network-Based Few-Shot Segmentation in Veterinary Medicine

  2024cites this paper
• Raman spectroscopic deep learning with signal aggregated representations for enhanced cell phenotype and signature identification

  2024cites this paper
• Label-free live cell recognition and tracking for biological discoveries and translational applications

  2024cites this paper
• Investigation of distributed learning for automated lesion detection in head MR images

  2024cites this paper
• Automating synaptic plasticity analysis: A deep learning approach to segmenting hippocampal field potential signal

  2024cites this paper
• Aerial Wildlife Image Repository for animal monitoring with drones in the age of artificial intelligence

  2024cites this paper
• A Multi-view Molecular Pre-training with Generative Contrastive Learning

  2024cites this paper
• Designing deep neural networks for driver intention recognition

  2024cites this paper
• Exploring the use of Synthetic Training Data for the Classification of Electronic Components in Artificial Intelligence Systems

  2024cites this paper
• SATDAUG - A Balanced and Augmented Dataset for Detecting Self-Admitted Technical Debt

  2024cites this paper
• Deep learning-based software bug classification

  2024cites this paper
• CBNN: 3-Party Secure Framework for Customized Binary Neural Networks Inference

  2024cites this paper
• AI-based Catalytic Performance Prediction for CO2 Electrochemical Reduction using Ionic Liquids

  2024cites this paper
• Hyperparameter tuning for deep learning semantic image segmentation of micro computed tomography scanned fiber-reinforced composites

  2024cites this paper
• Tourism and Hospitality Forecasting With Big Data: A Systematic Review of the Literature

  2024cites this paper
• How Many Data Does Machine Learning in Human–Computer Interaction Need?: Re-Estimating the Dataset Size for Convolutional Neural Network-Based Models of Visual Perception

  2023cites this paper
• Multi-Dataset Co-Training with Sharpness-Aware Optimization for Audio Anti-spoofing

  2023cites this paper
• Study on the real-time object detection approach for end-of-life battery-powered electronics in the waste of electrical and electronic equipment recycling process.

  2023cites this paper
• RoIA: Region of Interest Attention Network for Surface Defect Detection

  2023cites this paper
• Oriented object detection in optical remote sensing images using deep learning: a survey

  2023cites this paper
• Development of a Methane Emission Prediction Tool (POMEP178) for Palm Oil Mill Effluent Using Gaussian Process Regression

  2023cites this paper
• Automated Detection of Corneal Edema With Deep Learning-Assisted Second Harmonic Generation Microscopy

  2023cites this paper
• Advancements and Challenges in Machine Learning: A Comprehensive Review of Models, Libraries, Applications, and Algorithms

  2023cites this paper
• Artificial intelligence-aided method to detect uterine fibroids in ultrasound images: a retrospective study

  2023cites this paper
• Multitarget Intelligent Recognition of Petrographic Thin Section Images Based on Faster RCNN

  2023cites this paper
• On the training sample size and classification performance: An experimental evaluation in seismic facies classification

  2023cites this paper
• A power regulation strategy for heat pipe cooled reactors based on deep learning and hybrid data-driven optimization algorithm

  2023cites this paper
• Digital twin-driven vibration amplitude simulation for condition monitoring of axial blowers in blast furnace ironmaking

  2023cites this paper
• IIITH MM2 Speech-Text: A preliminary data for automatic spoken data validation with matched and mismatched speech-text content

  2023cites this paper
• Defectors: A Large, Diverse Python Dataset for Defect Prediction

  2023cites this paper
• The Magic Number: Impact of Sample Size for Dementia Screening Using Transfer Learning and Data Augmentation of Clock Drawing Test Images

  2023cites this paper
• A Reverse Auction-Based Incentive Mechanism for Cost-Effective Data Collection in Mobile Crowdsensing

  2023cites this paper
• A Wrapped Approach Using Unlabeled Data for Diabetic Retinopathy Diagnosis

  2023cites this paper
• Automated combustion model classification for char particle distributions using 3-D morphology analysis and pore-resolving CFD simulations

  2023cites this paper
• Classification of Alzheimer’s Disease using Random Oversampling and Albumentations on Convolutional Neural Network

  2023cites this paper
• Out-of-Distribution Data Generation for Fault Detection and Diagnosis in Industrial Systems

  2023cites this paper
• Advancing deep learning-based detection of floating litter using a novel open dataset

  2023cites this paper
• DNA-Storage in Future Communication Networks

  2023influential citation
• Quantitative Evaluation of a Multi-Modal Camera Setup for Fusing Event Data with RGB Images

  2023cites this paper
• Quantifying Privacy Risks of Prompts in Visual Prompt Learning

  2023cites this paper
• Generative Adversarial Network Models for Augmenting Digit and Character Datasets Embedded in Standard Markings on Ship Bodies

  2023cites this paper
• VisionKG: Unleashing the Power of Visual Datasets via Knowledge Graph

  2023cites this paper
• A Framework to Estimate the Key Point Within an Object Based on a Deep Learning Object Detection

  2023cites this paper
• Traffic congestion-aware graph-based vehicle rerouting framework from aerial imagery

  2023cites this paper
• Emotional Analysis Based On LSTM-CNN Hybrid Neural Network Model

  2023cites this paper
• Oriented Object Detection in Optical Remote Sensing Images: A Survey

  2023cites this paper
• Application of A Dual-Stage Deep Learning Framework to Detect Left Atrial Enlargement for Pet Heart Failure

  2023cites this paper
• Estimation of cyanobacteria pigments in the main rivers of South Korea using spatial attention convolutional neural network with hyperspectral imagery

  2022cites this paper
• Synthetic Training Image Dataset for Vision-Based 3D Pose Estimation of Construction Workers

  2022cites this paper
• ApacheJIT: A Large Dataset for Just-In-Time Defect Prediction

  2022cites this paper
• Self-adaptive temperature and humidity compensation based on improved deep BP neural network for NO2 detection in complex environment

  2022cites this paper
• Automated metadata annotation: What is and is not possible with machine learning

  2022cites this paper
• Prioritizing inspection of sewer pipes based on self-cleansing criteria

  2022cites this paper
• Fully Complex Deep Learning Classifiers for Signal Modulation Recognition in Non-Cooperative Environment

  2022cites this paper
• A real-time fingerprint-based indoor positioning using deep learning and preceding states

  2022cites this paper
• Paying attention for adjacent areas: Learning discriminative features for large-scale 3D scene segmentation

  2022cites this paper
• Low-Resource Adaptation for Personalized Co-Speech Gesture Generation

  2022cites this paper
• Brain age estimation using multi-feature-based networks

  2022cites this paper
• Language-Based Syllogistic Reasoning Using Deep Neural Networks

  2022cites this paper
• Performance of predictive supervised classification models of trace elements in magnetite for mineral exploration

  2022cites this paper
• Classification of myocardial fibrosis in DE-MRI based on semi-supervised semantic segmentation and dual attention mechanism

  2022cites this paper
• Deep Learning Model for Prediction of Entanglement Molecular Weight of Polymers

  2022cites this paper
• Managing Sustainability Tensions in Artificial Intelligence: Insights from Paradox Theory

  2022cites this paper
• Long-tailed Recognition by Learning from Latent Categories

  2022cites this paper
• Deep learning for behaviour classification in a preclinical brain injury model

  2022cites this paper
• Automated Defect Detection in Non-planar Objects Using Deep Learning Algorithms

  2022cites this paper
• DE-MRI myocardial fibrosis segmentation and classification model based on multi-scale self-supervision and transformer

  2022cites this paper
• Collate: Collaborative Neural Network Learning for Latency-Critical Edge Systems

  2022cites this paper
• MP-BADNet+\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$^+$$\end{document}: Secure and effective backdoor attack det

  2022cites this paper
• Synchronization-Enhanced Deep Learning Early Flood Risk Predictions: The Core of Data-Driven City Digital Twins for Climate Resilience Planning

  2022cites this paper
• Audio-Based Wildfire Detection on Embedded Systems

  2022cites this paper
• General intelligence requires rethinking exploration

  2022cites this paper
• DeepGuard: Backdoor Attack Detection and Identification Schemes in Privacy-Preserving Deep Neural Networks

  2022cites this paper
• A Comparative Study of Engraved-Digit Data Augmentation by Generative Adversarial Networks

  2022influential citation
• Learning Physically Meaningful Representations of Energy Systems with Variational Autoencoders

  2022cites this paper
• Trustworthiness of Laser-Induced Breakdown Spectroscopy Predictions via Simulation-based Synthetic Data Augmentation and Multitask Learning

  2022cites this paper