A parametric synthetic data generator for training learning-based sperm analysis systems

Daniel Hernández-Ferrándiz, Juan José Pantrigo, Raúl Cabido

Abstract: Highlights•We propose a parametric model of sperm adapted for different morphology classes.•We introduce the first synthetic sperm sample generator using this parametric model.•We demonstrate the efficacy of synthetic data for training supervised systems.•We test lightweight detection and tracking models to optimize the final system.