Multiple Classifier Combination Methodologies for Different Output Levels
MCS '00 Proceedings of the First International Workshop on Multiple Classifier Systems
Classifier Combinations: Implementations and Theoretical Issues
MCS '00 Proceedings of the First International Workshop on Multiple Classifier Systems
Confidence Evaluation for Combining Diverse Classifiers
ICDAR '03 Proceedings of the Seventh International Conference on Document Analysis and Recognition - Volume 2
Classifier combination based on confidence transformation
Pattern Recognition
| Hi-index | 0.00 |
This paper describes our investigation into the neural gas (NG) network algorithm and the hierarchical overlapped architecture (HONG) which have been built by retaining the essence of the original NG algorithm. By defining an implicit ranking scheme, the NG algorithm was made to run faster in its sequential implementation. The HONG network generated multiple classifications as confidence values, for every sample data presented. The final classification of the HONG architecture was obtained by combining these confidence values. Three HONG networks based on three different feature sets with global and structural features were also trained to obtain a better classification on handwritten data with high variations. An excellent recognition rate of 99.59% for the NIST SD3 database was consequently obtained.