#include <ApplyConcepts.h>
Collaboration diagram for Impala::Core::Training::ApplyConcepts:
Public Types | |
typedef Table::ScoreTable | ScoreTable |
typedef Feature::FeatureDefinition | FeatureDefinition |
typedef Feature::FeatureTable | FeatureTable |
typedef Feature::ConceptSet | ConceptSet |
Public Member Functions | |
ApplyConcepts (CmdOptions &options, Database::PathCreator *pathCreator) | |
c'tor reads the following from the command options: 1) read by mainVidSet or mainImSet 2) video/image-set that the models were trained on 3) filename of the concepts for which there are models 4) 'model type' 5) featuredef or precomputed kernel name | |
virtual | ~ApplyConcepts () |
virtual void | NextPath (Database::PathCreator *pathCreator) |
called from VideoSet:Listener::HandleNewFile or ImageSet::Listener::HandleNewFolder | |
Private Member Functions | |
ApplyConceptsHelper * | MakeHelper (bool isPrecomputed) |
Private Attributes | |
ApplyConceptsHelper * | mHelper |
ConceptSet * | mConceptSet |
Database::RawDataSet * | mAnnoSet |
ILOG_VAR_DEC |
It typically works with mainVidSet or mainImSet. Although the code is correct at the moment some code (and the interface) might need to change in the future. Especailly the use of factories can make the code a lot cleaner.
Definition at line 30 of file ApplyConcepts.h.