Matrix Classifier Trainer

Summary

Doc_BoxAlgorithm_MatrixClassifierTrainer.png
  • Plugin name : Matrix Classifier Trainer
  • Version : 0.1
  • Author : Thibaut Monseigne
  • Company : Inria
  • Short description : Matrix classifier trainer.
  • Documentation template generation date : Oct 29 2020

Description

Matrix classifier trainer.

Matrix classifier trainer. This box stack all matrix received in input and launch train function when a stimulation is received. Actual methods are Minimum Distance to Mean (MDM) and Minimum Distance to Mean with geodesic filtering (FgMDM) (With Real Time adaptation assumed) with or without Rebias

See also
CMatrixClassifierMDM::train, CMatrixClassifierMDMRebias::train, CMatrixClassifierFgMDMRT::train, CMatrixClassifierFgMDMRTRebias::train

Inputs

1. Stimulations

Stimulation to start the training.

  • Type identifier : Stimulations (0x6f752dd0, 0x082a321e)

2. Matrix for class 1

Input for Class 1

  • Type identifier : Streamed matrix (0x544a003e, 0x6dcba5f6)

3. Matrix for class 2

Input for Class 2

  • Type identifier : Streamed matrix (0x544a003e, 0x6dcba5f6)

Outputs

1. Tran-completed Flag

Send "OVTK_StimulationId_TrainCompleted" when train is completed.

  • Type identifier : Stimulations (0x6f752dd0, 0x082a321e)

Settings

1. Train trigger

Stimulation that starts the computation.

  • Type identifier : Stimulation (0x2c132d6e, 0x44ab0d97)
  • Default value : [ OVTK_StimulationId_Train ]

2. Filename to save classifier model

Classifier model Filename

  • Type identifier : Filename (0x330306dd, 0x74a95f98)
  • Default value : [ ${Player_ScenarioDirectory}/my-classifier.xml ]

3. Method

Classifier Method :
Minimum Distance to Mean (MDM)
Minimum Distance to Mean Rebias (MDM Rebias)
Minimum Distance to Mean with geodesic filtering (FgMDM)
Minimum Distance to Mean with geodesic filtering Rebias (FgMDM Rebias)

  • Type identifier : Matrix Classifier (0x5261636b, 0x436c6173)
  • Default value : [ Minimum Distance to Mean (MDM) ]

4. Metric

Log Level (None to see nothing)

  • Type identifier : Metric (0x5261636b, 0x4d455452)
  • Default value : [ Riemann ]

5. Log Level

Metric to use : Riemman, Euclidian, Harmonic, Identity, Kullback, Log Determinant, Log Euclidian

  • Type identifier : Log level (0xa88b3667, 0x0871638c)
  • Default value : [ Information ]

6. Class 1 label

Stimulation for Class 1.

  • Type identifier : Stimulation (0x2c132d6e, 0x44ab0d97)
  • Default value : [ OVTK_StimulationId_Label_01 ]

7. Class 2 label

Stimulation for Class 2.

  • Type identifier : Stimulation (0x2c132d6e, 0x44ab0d97)
  • Default value : [ OVTK_StimulationId_Label_02 ]

Examples

Miscellaneous