Fetching latest commit…
Cannot retrieve the latest commit at this time.
|Failed to load latest commit information.|
RECOGNITION MODULE Author: Octavian Sima It recognizes persons provided by the detection module using the Eigenfaces algorithm. The module is very flexible and all the available options are detailed in the Recognition.h file. To take advantage of the OOP style, the module implements the IRecognition interface, same as the Detection module implements the IDetection interface. The main options are set by calling the "initialize" function. If you ommit this, default options will be used. Before recognizing a face, you must call the "train" function. Its parameters are explained in Recognition.h. Training data is stored in output files as images and xml. HOWTO USE: 1. instantiate the recognition Class: Recognition* recognition = new Recognition(); 2. call recognition->initialize(...); 3. perform training : recognition->train(...); 4(optional) save data: recognition->saveTrainingData(TRAINING_DATA_FILE); to release memory. When you need training data, call recognition->loadTrainingData(...)/ 5. to recognize new faces, call recognition->recognizeFaces(...) To test recognition performance, use the Test.cpp file. It calls recognition->testRecognitionPerformance(...) for different test files from "tests" folder. For the default test use the next steps: 1. make 2. make train 3. make test for other options, call: ./recognition and the following text will be shown: Usage: For training application with default options run as ./recognition train For training application from database (such person_database) run as ./recognition train database databaseFolderPath For training application from a file (ex in tests/ ) run as ./recognition train file trainFilePath For testing application with default options run as ./recognition test For testing application from a file (ex in tests/ ) run as ./recognition test testFilePath recognitionThreshold For help run as ./recognition help