Classifier creation framework for diverse classification tasks

Classifier creation framework for diverse classification tasks

Software

This software package contains the extendable framework for classifiers creation of various tasks. The only requirements for input object type is ability to be converted to set of feature vectors, the rest of computing is done by framework, finally the classifier is created. The framework can be used for clasifiers creation using training part of dataset and then do the testing on testing part and thus show the classifier accuracy. The framework consists of several basic parts which are replaceable such as feature extractor, codebook generator, codebook translator and classification technique. Sample applications of these are included. The whole process can be computed on local computer or in Sun Grid Engine schedulling system. The software package contains the framework files, sample feature extractor, sample classification technique and the particular documentation of system setup and the user guide.

This software package contains the extendable framework for classifiers creation of various tasks. The only requirements for input object type is ability to be converted to set of feature vectors, the rest of computing is done by framework, finally the classifier is created. The framework can be used for clasifiers creation using training part of dataset and then do the testing on testing part and thus show the classifier accuracy. The framework consists of several basic parts which are replaceable such as feature extractor, codebook generator, codebook translator and classification technique. Sample applications of these are included. The whole process can be computed on local computer or in Sun Grid Engine schedulling system. The software package contains the framework files, sample feature extractor, sample classification technique and the particular documentation of system setup and the user guide.

framework, feature extractors, linear SVM, SGE

framework, feature extractors, linear SVM, SGE

http://www.fit.vutbr.cz/research/prod/index.php.cs?id=155

výsledek využívá pouze poskytovatel

Využití výsledku jiným subjektem je možné bez nabytí licence (výsledek není licencován)

http://www.fit.vutbr.cz/research/prod/index.php.cs?id=155

framework