Використання gan-моделі для створення даних у задачах детектування дронів

Abstract

This work addresses the demand for sophisticated techniques in drone detec-tion in the context of rapidly evolving technologies. Emphasis is placed on the utilization of Generative Adversarial Networks (GANs) to generate artificial au-dio data of drones, which possesses the potential to substantially enhance and di-versify the datasets for training machine learning models. Through the deploy-ment of GAN-generated audio samples, researchers and practitioners can mitigate the costs and challenges associated with acquiring large-scale real-world data, thereby improving the efficacy of existing detection models.