Constructor

obj = cv.FacemarkKazemi()
obj = cv.FacemarkKazemi('OptionName',optionValue, ...)

Options

• CascadeDepth depth of cascade used for training. default 15
• TreeDepth max height of the regression tree built. default 5
• NumTreesPerCascadeLevel number of trees fit per cascade level. default 500
• LearningRate learning rate in gradient boosting, also refered to as shrinkage. default 0.1
• OversamplingAmount number of initializations used to create training samples. default 20
• NumTestCoordinates number of test coordinates. default 500
• Lambda value to calculate probability of closeness of two coordinates. default 0.1
• NumTestSplits number of random test splits generated. default 20
• ConfigFile name of the file containing the values of training parameters. default ''

These variables are used for training data. They are initialised as described in the referenced research paper.