Deep learning-based framework for expansion, recognition and classification of underwater acoustic signal

作者：Guanghao Jin, Fan Liu, Hao Wu, Qingzeng Song · 发表于：Journal of Experimental & Theoretical Artificial Intelligence · 年份：2019 · DOI：10.1080/0952813x.2019.1647560 · 被引用次数：73 · 研究领域：Underwater Acoustics Research、Speech and Audio Processing、Blind Source Separation Techniques

Recently, deep learning has developed rapidly and contributed in many fields like the classification in radar and sonar applications. In some special fields like the underwater acoustic signals, the dataset for training may be scarce due to the reason of security or other restrictions, which affects the performance of the deep learning methods as those need a big dataset to ensure high accuracy. Furthermore, the original dataset is in some formats like audio, which makes those methods difficult to capture features, especially in insufficient sample case because of the interference. In this paper, we present a novel framework that applies the LOFAR spectrum for preprocessing to retain key features and utilises Generative Adversarial Networks (GAN) for the expansion of samples to improve the performance classification. Firstly, our framework selects proper preprocessing method based on the evaluation of the spectrum methods. Secondly, our framework revises a GAN to generate samples and built an independent classification network to ensure the quality of those. Finally, our framework applies the existing classification networks to evaluate the performance and selects the best one for real utilisation. The experimental results show that the generated samples have high quality, which can significantly improve the classification accuracy of the neural models.