Multiple Classifier Combination Methodologies for Different Output Levels
Lecture notes in computer sciencePublished 1 January 2000
Ching Y. Suen, Louisa Lam
Citations121
SJR quartileQ2
SJR score0.35
SNIP0.55
Generate an AI Snapshot to get a quick, structured summary of this paper.
Study Snapshot
ObjectiveStudy objective
MethodsResearch methodology
PopulationPopulation studied
Sample sizeSample sizes
OutcomesStudy outcomes here
ResultsStudy results comes here
LimitationsResearch study limitations comes here
A concise AI-generated summary of the paper will appear here once you click Generate AI Snapshot.
TL;DR
This article examines the main combination methods that have been developed for different levels of classifier outputs - abstract level, ranked list of classes, and measurements.
Abstract
In the past decade, many researchers have employed various methodologies to combine decisions of multiple classifiers in order to order to improve recognition results. In this article, we will examine the main combination methods that have been developed for different levels of classifier outputs - abstract level, ranked list of classes, and measurements. At the same time, various issues, results, and applications of these methods will also be considered, and these will illustrate the diversity and scope of this research area.
Keywords
ChemistryComputer Science
IEEE Transactions on Pattern Analysis and Machine IntelligenceThe random subspace method for constructing decision forests
6,737 Citations1998Tin Kam Ho
A method to construct a decision tree based classifier is proposed that maintains highest accuracy on training data and improves on generalization accuracy as it grows in complexity.
IEEE Transactions on Pattern Analysis and Machine IntelligenceOn combining classifiers
5,227 Citations1998Josef Kittler, M. Hatef +2 more
IEEE Transactions on Systems Man and CyberneticsMethods of combining multiple classifiers and their applications to handwriting recognition
2,171 Citations1992Longchang Xu, Adam Krzyżak +1 more
On applying these methods to combine several classifiers for recognizing totally unconstrained handwritten numerals, the experimental results show that the performance of individual classifiers can be improved significantly.
IEEE Transactions on Pattern Analysis and Machine IntelligenceDecision combination in multiple classifier systems
1,511 Citations1994Tin Kam Ho, J.J. Hull +1 more
This work proposes three methods based on the highest rank, the Borda count, and logistic regression for class set reranking that have been tested in applications of degraded machine-printed characters and works from large lexicons, resulting in substantial improvement in overall correctness.
IEEE Transactions on Systems Man and Cybernetics - Part A Systems and HumansApplication of majority voting to pattern recognition: an analysis of its behavior and performance
842 Citations1997Louisa Lam, Shu-Ping Suen
This paper examines the mode of operation of the majority vote method in order to gain a deeper understanding of how and why it works, so that a more solid basis can be provided for its future applications to different data and/or domains.
IEEE Transactions on Pattern Analysis and Machine IntelligenceA method of combining multiple experts for the recognition of unconstrained handwritten numerals
561 Citations1995Yongliang Huang, Ching Y. Suen
A method was developed which can aggregate the decisions obtained from individual classifiers and derive the best final decisions from the statistical point of view, and outperforms voting, Bayesian, and Dempster-Shafer approaches.
Pattern Analysis and ApplicationsCombining classifiers: A theoretical framework
375 Citations1998Josef Kittler
Classifier combination can be viewed as a multistage classification process whereby thea posteriori class probabilities generated by the individual classifiers are considered as features for a second stage classification scheme.
Pattern Recognition LettersOptimal combinations of pattern classifiers
309 Citations1995Louisa Lam, Ching Y. Suen
A Bayesian formulation and a weighted majority vote (with weights obtained through a genetic algorithm) are implemented, and the combined performances of 7 classifiers on a large set of handwritten numerals are analyzed.
Pattern Recognition LettersCombining multiple matchers for a high security fingerprint verification system
209 Citations1999Anil K. Jain, Salil Prabhakar +1 more
This work uses the logistic transform to integrate the output scores from three different fingerprint matching algorithms and confirms the effectiveness of the proposed integration scheme on a large fingerprint database.
Combining classifiers
198 Citations1996Josef Kittler, Mary Ann Hater +1 more
Pattern Recognition LettersExperimental evaluation of expert fusion strategies
118 Citations1999Fuad M. Alkoot, Josef Kittler
A relation between the effectiveness of various combination strategies and estimation errors is established and the behaviour of individual experts is emulated by subjecting their nominal soft outputs to perturbation errors.
Machine intelligence and pattern recognitionCombining the Classification Results of Independent Classifiers Based on the Dempster/Shafer Theory of Evidence
104 Citations1988E. Mandler, Jürgen Schumann
A procedure is defined that transform the distance measures of different classifiers into confidence values in three subsequent steps and achieves an output vector for each distance classifier which is compatible with statistically adapted classifiers.
Pattern Analysis and ApplicationsReliability Parameters to Improve Combination Strategies in Multi-Expert Systems
86 Citations1999L.P. Cordella, Pasquale Foggia +3 more
A method for estimating the reliability of each single recognition act of an expert on the basis of information directly derived from its output is proposed, allowing a more precise weighting during the combination phase of recognition systems.
IEEE Transactions on Systems Man and Cybernetics Part B (Cybernetics)Use of fuzzy-logic-inspired features to improve bacterial recognition through classifier fusion
55 Citations1998Dayou Wang, J.M. Keller +3 more
The fuzzy integral was utilized in the fusion of neural networks trained with different feature sets to reach an almost perfect classification rate of E. coli O157:H7 PFGE patterns made available for the experiments.
Pattern RecognitionAN EVALUATION OF MULTI-EXPERT CONFIGURATIONS FOR THE RECOGNITION OF HANDWRITTEN NUMERALS
49 Citations1998Fuad Rahman, M.C. Fairhurst
A new class of decision combination approaches is presented and the effectiveness of these approaches in successfully combining decisions by multiple experts in the specific application of handwritten numeral recognition is compared.
Pattern Recognition LettersOptimal approximation of discrete probability distribution with kth-order dependency and its application to combining multiple classifiers
47 Citations1997Hee-Joong Kang, Kawon Kim +1 more
A new method is proposed to optimally approximate a high order distribution with a product of kth-order dependencies, or (k + 1)st-order distributions, where 1 ⩽ k⩽ K.
Pattern Analysis and ApplicationsSerial Combination of Multiple Experts: A Unified Evaluation
44 Citations1999Fuad Rahman, M.C. Fairhurst
A unified framework for serial multiple expert decision combination is presented, showing that many multi-expert approaches reported in the literature can be easily represented within the proposed framework.
HMM-KNN word recognition engine for bank cheque processing
41 Citations2002D. Guillevic, C.Y. Suen
The mixed HMM-KNN word recognition module of a bank cheque processing system developed at CENPARMI is described, which can be easily adapted to read other European languages based on the Roman alphabet.
A methodology of combining HMM and MLP classifiers for cursive word recognition
24 Citations2002Jin Ho Kim, Kye Kyung Kim +2 more
A methodology of combining HMM (hidden Markov model) and MLP (multilayer perceptron) for cursive word recognition is presented and a new probability measure for the hybrid classifier as well as conventional combining schemes are introduced.
Pattern Analysis and ApplicationsDual classifier system for handprinted alphanumeric character recognition
23 Citations1998Yeun-Ting Chim, Ashraf A. Kassim +1 more
This paper presents a dual classifier handprinted character recognition system that is implemented using Radial Basis Function (RBF) networks, and the results are fourd to be very promising.
Writing style detection by statistical combination of classifiers in form reader applications
22 Citations2002J. Franke, Matthias Oberländer
The authors deal with the recognition of writing style (whether a data field is hand or machine printed) in the context of form reading applications by developing different statistical classifiers which were adapted to different feature vectors calculated separately for each data field.
Combining classifiers based on minimization of a Bayes error rate
21 Citations1999Heejoong Kang, Seong–Whan Lee
Multiple classifiers recognizing unconstrained handwritten numerals were combined by the proposed approximation scheme based on the minimization of the Bayes error rate, and the high recognition rates were obtained by them.
Combining classifiers based on confidence values
19 Citations1999Ajantha S. Atukorale, Ponnuthurai Nagaratnam Suganthan
The paper describes the investigation into the neural gas (NG) network algorithm and the hierarchical overlapped architecture (HONG) which was built by retaining the essence of the original NG algorithm by defining an implicit ranking scheme.
Series in machine perception and artificial intelligenceADVANCES IN DOCUMENT CLASSIFICATION BY VOTING OF COMPETITIVE APPROACHES
14 Citations1998C. Wenzel, S. Baumann +1 more
This paper presents a complex approach for the content-based text categorization of printed German business letters into pre-defined message types such as order, invoice, offer, etc.
Evaluation of combination methods
13 Citations1999S. Impedovo, A. Salzo
A new approach to evaluate the performances of combination methods is presented, which takes into account both the recognition rates of the experts combined and the correlation among them and allows the selection of the best combination method for each set of experts.
International Journal of Pattern Recognition and Artificial IntelligenceDifferentiating Between Oriental and European Scripts by Statistical Features
12 Citations1998Louisa Lam, Jie Ding +1 more
This paper presents a method which uses a combined analysis of several discriminating statistical features, for the differentiation between European and oriental language scripts, which has proved to be effective in differentiating between documents printed in these different scripts.
A dependency-based framework of combining multiple experts for the recognition of unconstrained handwritten numerals
9 Citations2003Hee-Joong Kang, Seong‐Whan Lee
A dependency-based framework is proposed to optimality approximate a probability distribution with a product set of dth-order dependencies where 1<d<K, and to combine multiple experts based on the product set using the Bayesian formalism.
International Conference on Advances in Pattern RecognitionPattern Classification: Fusion of Information
9 Citations1999Josef Kittler
It is shown that the use of multiple cues can be formulated as a problem of fusing the opinions of multiple experts employing distinct representations, which contrasts with the methods aimed at improving estimates of decision function values.
MergeLayouts-overcoming faulty segmentations by a comprehensive voting of commercial OCR devices
6 Citations1999Stefan Klink, T. Jager
A comprehensive voting approach, taking entire layouts obtained from commercial OCR devices as input, which achieves a "better" layout by combining all attributes of a segment, representing the original page layout as good as possible.