Home

Awesome

Gesture Recognition Toolkit (GRT)

The Gesture Recognition Toolkit (GRT) is a cross-platform, open-source, C++ machine learning library designed for real-time gesture recognition.

Build Status:

Key things to know about the GRT:

##Core Resources

##GRT Architecture To support flexibility while maintaining consistency, the GRT uses an object-oriented modular architecture. This architecture is built around a set of core modules and a central gesture recognition pipeline.

The input to both the modules and pipeline consists of an N-dimensional double-precision vector, making the toolkit flexible to the type of input signal. The algorithms in each module can be used as standalone classes; alternatively a pipeline can be used to chain modules together to create a more sophisticated gesture-recognition system. The GRT includes modules for preprocessing, feature extraction, clustering, classification, regression and post processing.

The toolkit's source code is structured as following:

##Getting Started Example This example demonstrates a few key components of the GRT, such as:

You can find this source code and a large number of other examples and tutorials in the GRT examples folder.

//Include the main GRT header
#include <GRT/GRT.h>
using namespace GRT;

int main (int argc, const char * argv[])
{
    //Generate a basic dummy dataset with 1000 samples, 5 classes, and 3 dimensions
    cout << "Generating dataset..." << endl;
    ClassificationData::generateGaussDataset( "data.csv", 1000, 5, 3 );
	
    //Load some training data from a file
    ClassificationData trainingData;

    cout << "Loading dataset..." << endl;
    if( !trainingData.load( "data.csv" ) ){
		cout << "ERROR: Failed to load training data from file\n";
		return EXIT_FAILURE;
    }

    cout << "Data Loaded" << endl;

    //Print out some stats about the training data
    trainingData.printStats();

    //Partition the training data into a training dataset and a test dataset. 80 means that 80%
    //of the data will be used for the training data and 20% will be returned as the test dataset
    cout << "Splitting data into training/test split..." << endl;
    ClassificationData testData = trainingData.partition(80);

    //Create a new Gesture Recognition Pipeline using an Adaptive Naive Bayes Classifier
    GestureRecognitionPipeline pipeline;
    pipeline.setClassifier( ANBC() );

    //Train the pipeline using the training data
    cout << "Training model..." << endl;
    if( !pipeline.train( trainingData ) ){
        cout << "ERROR: Failed to train the pipeline!\n";
        return EXIT_FAILURE;
    }

    //Save the pipeline to a file
    if( !pipeline.save( "HelloWorldPipeline" ) ){
        cout << "ERROR: Failed to save the pipeline!\n";
        return EXIT_FAILURE;
    }

    //Load the pipeline from a file
    if( !pipeline.load( "HelloWorldPipeline" ) ){
        cout << "ERROR: Failed to load the pipeline!\n";
        return EXIT_FAILURE;
    }

    //Test the pipeline using the test data
    cout << "Testing model..." << endl;
    if( !pipeline.test( testData ) ){
        cout << "ERROR: Failed to test the pipeline!\n";
        return EXIT_FAILURE;
    }

    //Print some stats about the testing
    cout << "Test Accuracy: " << pipeline.getTestAccuracy() << endl;
   
    vector< UINT > classLabels = pipeline.getClassLabels();

    cout << "Precision: ";
    for(UINT k=0; k<pipeline.getNumClassesInModel(); k++){
        cout << "\t" << pipeline.getTestPrecision( classLabels[k] );
    }cout << endl;

    cout << "Recall: ";
    for(UINT k=0; k<pipeline.getNumClassesInModel(); k++){
         cout << "\t" << pipeline.getTestRecall( classLabels[k] );
    }cout << endl;

    cout << "FMeasure: ";
    for(UINT k=0; k<pipeline.getNumClassesInModel(); k++){
        cout << "\t" << pipeline.getTestFMeasure( classLabels[k] );
    }cout << endl;

    MatrixDouble confusionMatrix = pipeline.getTestConfusionMatrix();
    cout << "ConfusionMatrix: \n";
    for(UINT i=0; i<confusionMatrix.getNumRows(); i++){
        for(UINT j=0; j<confusionMatrix.getNumCols(); j++){
            cout << confusionMatrix[i][j] << "\t";
        }cout << endl;
    }

    return EXIT_SUCCESS;
}

##Tutorials and Examples

You can find a large number of tutorials and examples in the examples folder. You can also find a wide range of examples and references on the main GRT wiki:

http://www.nickgillian.com/wiki/pmwiki.php?n=GRT.GestureRecognitionToolkit

If you build the GRT using CMake, an examples folder will automatically be generated in the build directory after you successfully build the main GRT library. Example applications can then be directly run from this example directory. To run any of the examples, open terminal in the grt/build/examples directory and run:

./ExampleName

where ExampleName is the name of the example application you want to run.

##Forum

You can find the main GRT forum at: http://www.nickgillian.com/forum/

##Building the GRT

You can find a CMakeLists file in the build folder that you can use to autogenerate a makefile for your machine.

Read the readme file in the build folder to see how to build the GRT as a static library for Linux, OS X, or Windows.

##Installing and using the GRT in your C++ projects

See the build directory for details on how to build, install, and use the GRT in your C++ projects.

##Android port of the Gesture Recognition Toolkit

as described [http://hollyhook.de/wp/grt-for-android/] extended and fixed

##License

The Gesture Recognition Toolkit is available under a MIT license.

Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:

The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software. THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.