Using generative models for handwritten digit recognition

Revow, Michael; Williams, Christopher K. I.; Hinton, Geoffrey E.

doi:10.1109/34.506410

Cited by 152 publications

(91 citation statements)

References 40 publications

(17 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Secondly, the model is made more compact by eliminating some of the variability caused by control points shifting along the contour (which cause little change to the actual observed shape). This work has some similarity to the work of Revow et al [9] in that a covariance matrix associated with control point positions is learned from training data using an iterative learning process.…”

Section: Introductionmentioning

confidence: 99%

An Adaptive Eigenshape Model.

Baumberg

Hogg²

1995

Procedings of the British Machine Vision Conference 1995

View full text Add to dashboard Cite

There has been a great deal of recent interest in statistical models of 2D landmark data for generating compact deformable models of a given object. This paper extends this work to a class of parametrised shapes where there are no landmarks available. A rigorous statistical framework for the eigenshape model is introduced, which is an extension to the conventional Linear Point Distribution Model.One of the problems associated with landmark free methods is that a large degree of variability in any shape descriptor may be due to the choice of parametrisation. An automated training method is described which utilises an iterative feedback method to overcome this problem. The result is an automatically generated compact linear shape model. The model has been successfully applied to a problem of tracking the outline of a walking pedestrian in real time.

show abstract

Section: Introductionmentioning

confidence: 99%

An Adaptive Eigenshape Model.

Baumberg

Hogg²

1995

Procedings of the British Machine Vision Conference 1995

View full text Add to dashboard Cite

show abstract

“…The aggregation mechanism is a vital part of each majority vote model. Revow et al compare five combination strategies (majority vote, Bayesian, logistic regression, fuzzy integral, and neural network) and arrive at the conclusion that majority vote is as effective as the other, more complicated schemes to improve the recognition rate for the data set used [18].…”

Section: Related Workmentioning

confidence: 99%

Statistical Quality Control for Human-Based Electronic Services

Kern

Thies

Satzger

2010

Service-Oriented Computing – ICSOC 2007

View full text Add to dashboard Cite

Crowdsourcing in form of human-based electronic services (people services) provides a powerful way of outsourcing tasks to a large crowd of remote workers over the Internet. Research has shown that multiple redundant results delivered by different workers can be aggregated in order to achieve a reliable result. However, existing implementations of this approach are rather inefficient as they multiply the effort for task execution and are not able to guarantee a certain quality level. As a starting point towards an integrated approach for quality management of people services we have developed a quality management model that combines elements of statistical quality control (SQC) with group decision theory. The contributions of the workers are tracked and weighted individually in order to minimize the quality management effort while guaranteing a well-defined level of overall result quality. A quantitative analysis of the approach based on an optical character recognition (OCR) scenario confirms the efficiency and reach of the approach.

show abstract

“…Revow et al [14] developed a model for characters by hand-specifying and then adapting a set of "control points," defining the probability of an observed character as a function of the deviation of the character from these predefined control points. The authors discounted the affine component of such deviations, achieving an affine 11.…”

Section: Additional Related Workmentioning

confidence: 99%

Data driven image models through continuous joint alignment

Learned-Miller

2006

IEEE Trans. Pattern Anal. Machine Intell.

265

185

View full text Add to dashboard Cite

Abstract-This paper presents a family of techniques that we call congealing for modeling image classes from data. The idea is to start with a set of images and make them appear as similar as possible by removing variability along the known axes of variation. This technique can be used to eliminate "nuisance" variables such as affine deformations from handwritten digits or unwanted bias fields from magnetic resonance images. In addition to separating and modeling the latent images-i.e., the images without the nuisance variables-we can model the nuisance variables themselves, leading to factorized generative image models. When nuisance variable distributions are shared between classes, one can share the knowledge learned in one task with another task, leading to efficient learning. We demonstrate this process by building a handwritten digit classifier from just a single example of each class. In addition to applications in handwritten character recognition, we describe in detail the application of bias removal from magnetic resonance images. Unlike previous methods, we use a separate, nonparametric model for the intensity values at each pixel. This allows us to leverage the data from the MR images of different patients to remove bias from each other. Only very weak assumptions are made about the distributions of intensity values in the images. In addition to the digit and MR applications, we discuss a number of other uses of congealing and describe experiments about the robustness and consistency of the method.

show abstract

Using generative models for handwritten digit recognition

Cited by 152 publications

References 40 publications

An Adaptive Eigenshape Model.

An Adaptive Eigenshape Model.

Statistical Quality Control for Human-Based Electronic Services

Data driven image models through continuous joint alignment

Contact Info

Product

Resources

About