Papperoni

2006

One-Shot Learning of Object Categories

Li Fei Fei, Rob Fergus, Pietro Perona

Open PDF Google Scholar

citations

Cite Score

66

AI summary

This paper introduces a Bayesian one-shot learning algorithm for object categories, utilizing prior knowledge from previously learned categories to learn new categories from one or a few images, achieving informative models on a database of 101 diverse object categories.

Main Contributions

Proposed a Bayesian approach for one-shot learning of object categories, leveraging prior knowledge from previously learned categories.
Utilized a Bayesian implementation where prior knowledge is represented as a probability density function on model parameters, updated with new observations.
Evaluated the algorithm on a database of 101 diverse object categories, demonstrating its effectiveness with a limited number of training examples.
Compared the Bayesian approach to Maximum Likelihood (ML) and Maximum A Posteriori (MAP) methods, showing superior performance when training data is scarce.
Demonstrated that useful information about object categories can be obtained from very few, even one, training example.

Abstract

Learning visual models of object categories notoriously requires hundreds or thousands of training examples. We show that it is possible to learn much information about a category from just one, or a handful, of images. The key insight is that, rather than learning from scratch, one can take advantage of knowledge coming from previously learned categories, no matter how different these categories might be. We explore a Bayesian implementation of this idea. Object categories are represented by probabilistic models. Prior knowledge is represented as a probability density function on the parameters of these models. The posterior model for an object category is obtained by updating the prior in the light of one or more observations. We test a simple implementation of our algorithm on a database of 101 diverse object categories. We compare category models learned by an implementation of our Bayesian approach to models learned from by Maximum Likelihood (ML) and Maximum A Posteriori (MAP) methods. We find that on a database of more than 100 categories, the Bayesian approach produces informative models when the number of training examples is too small for other methods to operate successfully.

Citation Graph

Loading graph...

References [43]

Sort:

Filter:

[1]Gradient-Based Learning Applied to Document Recognition

Yann Lecun, Leon Bottou, Yoshua Bengio, Patrick Haffner - 1998

62 papers in library cite

I absolutely hated this paper. Has ~50 pages but seems like 200 pages. Takes too long to explain some things that really is just repeating itself. Also doesn't seem to add too much on top of LeNet-5. Also, focuses a lot on GTNs, which really didn't stick.

[2]Learning Generative Visual Models From Few Training Examples: An Incremental Bayesian Approach Tested on 101 Object Categories

Li Fei Fei, Rob Fergus, Pietro Perona - 2004

15 papers in library cite

I think most people cite this thinking this is where the Caltech 101 dataset comes from (it's not). Anyway, it's just an extension of the other dataset and it's very mathy, not NNs, and uninteresting.

[3]Neural Network-Based Face Detection

H. Rowley, S. Baluja, Takeo Kanade - 1998

4 papers in library cite

It's nice, but nothing crazy. Just applying CNNs

[4]Learning Methods for Generic Object Recognition With Invariance to Pose and Lighting

Yann Lecun, Fu Jie Huang, Leon Bottou - 2004

18 papers in library cite

Good paper, nice methodology for creating different images. However, I think that this was not too impactful... I don't see this being used a lot.

[5]Rapid Object Detection Using a Boosted Cascade of Simple Features

P. Viola, M. J. Jones - 2001

10 papers in library cite

[6]Shape Matching and Object Recognition Using Low Distortion Correspondences

A. C. Berg, T. L. Berg, Jitendra Malik - 2005

8 papers in library cite

[7]Maximum Likelihood From Incomplete Data via the EM Algorithm

A. P. Dempster, N. M. Laird, D. B. Rubin - 1977

6 papers in library cite

[8]Object Recognition From Local Scale-Invariant Features

D. G. Lowe - 1999

6 papers in library cite

[9]A Bayesian Approach to Unsupervised One-Shot Learning of Object Categories

Li Fei Fei, Rob Fergus, Pietro Perona - 2003

4 papers in library cite

[10]A New View of the EM Algorithm That Justifies Incremental, Sparse and Other Variants

Radford M. Neal, Geoffrey E. Hinton - 1998

4 papers in library cite

[11]Object Class Recognition by Unsupervised Scale-Invariant Learning

Rob Fergus, Pietro Perona, Andrew Zisserman - 2003

4 papers in library cite

[12]Sharing Features: Efficient Boosting Procedures for Multiclass Object Detection

Antonio Torralba, Kevin P. Murphy, William T. Freeman - 2004

4 papers in library cite

[13]A Computational Model for Visual Selection

Y. Amit, D. Geman - 1999

3 papers in library cite

[14]A Statistical Method for 3d Object Detection Applied to Faces and Cars

H. Schneiderman, Takeo Kanade - 2000

3 papers in library cite

[15]A Visual Category Filter for Google Images

Rob Fergus, Pietro Perona, Andrew Zisserman - 2004

3 papers in library cite

[16]Example-Based Learning for View-Based Human Face Detection

K. Sung, T. Poggio - 1994

3 papers in library cite

[17]Recognition-by-Components: A Theory of Human Image Understanding

I. Biederman - 1987

3 papers in library cite

[18]Scale, Saliency and Image Description

T. Kadir, M. Brady - 2001

3 papers in library cite

[19]Unsupervised Learning of Models for Recognition

M. Weber, M. Welling, Pietro Perona - 2000

3 papers in library cite

[20]Unsupervised Learning of Models for Recognition

M. Weber, M. Welling, Pietro Perona - 2000

3 papers in library cite

[21]A Probabilistic Approach to Object Recognition Using Local Photometry and Global Geometry

M. Burl, M. Weber, Pietro Perona - 1996

2 papers in library cite

[22]An Affine Invariant Interest Point Detector

K. Mikolajczyk, Cordelia Schmid - 2002

2 papers in library cite

[23]Inferring Parameters and Structure of Latent Variable Models by Variational Bayes

H. Attias - 1999

2 papers in library cite

Missing author list

[24]Merriam-Webster's Collegiate Dictionary

1994

2 papers in library cite

[25]Variational Bayes for D-Dimensional Gaussian Mixture Models

W. Penny - 2001

2 papers in library cite

[26]Visual Object Category Recognition

Rob Fergus - 2005

2 papers in library cite

[27]3D Object Modeling and Recognition Using Affine-Invariant Patches and Multi-View Spatial Constraints

F. Rothganger, Svetlana Lazebnik, Cordelia Schmid, Jean Ponce - 2003

1 paper in library cites

[28]A Sparse Object Category Model for Efficient Learning and Exhaustive Recognition

Rob Fergus, Pietro Perona, Andrew Zisserman - 2005

1 paper in library cites

[29]Adaptive Rejection Sampling for Gibbs Sampling

R. Gilks, P. Wild - 1992

1 paper in library cites

[30]Bayesian Data Analysis

A. Gelman, J. B. Carlin, H. S. Stern, D. B. Rubin - 1995

1 paper in library cites

[31]Combining Class-Specific Fragments for Object Classification

E. Sali, S. Ullman - 1999

1 paper in library cites

[32]Comparing Images Using the Hausdorff Distance

D. Huttenlocher, G. Klanderman, W. Rucklidge - 1993

1 paper in library cites

[33]Detecting Pedestrians Using Patterns of Motion and Appearance

P. Viola, M. J. Jones, D. Snow - 2003

1 paper in library cites

[34]Finding Faces in Cluttered Scenes Using Labeled Random Graph Matching

T. Leung, M. Burl, Pietro Perona - 1995

1 paper in library cites

[35]Markov Chain Monte Carlo in Practice

R. Gilks, S. Richardson, D. Spiegelhalter - 1992

1 paper in library cites

[36]On the Sensitivity of the Hough Transform for Object Recognition

W. Grimson, D. Huttenlocher - 1990

1 paper in library cites

[37]Pictorial Structures for Object Recognition

P. F. Felzenszwalb, D. Huttenlocher - 2005

1 paper in library cites

[38]Recognition of Planar Object Classes

M. Burl, Pietro Perona - 1996

1 paper in library cites

[39]Representation and Detection of Deformable Shapes

P. F. Felzenszwalb, D. Huttenlocher - 2005

1 paper in library cites

[40]Shape From Shading in the Light of Mutual Illumination

David Forsyth, Andrew Zisserman - 1990

1 paper in library cites

[41]Some Examples of Recursive Variational Approximations for Bayesian Inference

K. Humphreys, M. Titterington - 2001

1 paper in library cites

[42]Supplemental Material

Li Fei Fei, Rob Fergus, Pietro Perona - 2006

1 paper in library cites

[43]Viewpoint-Invariant Learning and Detection of Human Heads

M. Weber, W. Einhaeuser, M. Welling, Pietro Perona - 2000

1 paper in library cites

Cited by

5

papers in your library

Cites

4

papers in your library

Read

on January 25, 2026

It's a nice problem they are trying to solve, though I doubt the effectiveness. Maybe it was really relevant in a time where they didn't have data. Problem is it focuses too much on the methodology, and I am not too interested in it.

Tags

Paper Aliases

No aliases