Zero shot objects classification method of side scan sonar image based on synthesis of pseudo samples. (February 2021)

• Record Type: Journal Article
• Authors: Li, Chuanlong; Ye, Xiufen; Cao, Dongxiang; Hou, Jie; Yang, Haibo
• Is Part Of: Applied acoustics. Volume 173(2021)
• Publication Date: 2021-02
• Subjects: Zero shot learning -- Pseudo-sample synthesis -- Style transfer -- Deep learning -- Side-scan sonar image classification; Acoustical engineering -- Periodicals; Periodicals; 620.2
• DOI: 10.1016/j.apacoust.2020.107691
• Languages: English
• ISSNs: 0003-682X

Abstract

Graphical abstract: Highlights: A zero-shot side-scan sonar image targets classification method is proposed. The proposed method does not need real labeled side-scan sonar samples. The proposed method can synthesize pseudo-samples for any categories. Experimental results on our side-scan sonar image dataset verified the performance of the our method. Abstract: Side-scan sonar (SSS) is the most important sensor in the field of ocean exploration, auto-detection of targets in SSS images is very critical. Deep neural networks (DNNs) have exhibited impressive performance but require very large number of training samples; only a limited number of SSS images will not suffice. In more extreme cases, no appropriate SSS image is available for some specific targets that need to be identified, making it impossible to train DNNs. In this paper, we deal with such extreme situations, how can a DNN recognize targets in SSS images without any training samples? which is the "zero-shot learning problem." Inspired by the way humans perceive the world, we develop a zero-shot SSS image classification method through synthesis of pseudo SSS images. For a given category, we use a fixed style-transfer method to synthesize pseudo samples using common optical images and any available SSS images, and train the DNN with these pseudo samples. The zero-shot learning problem thus can be transformed to a conventional supervised learning problem. Experimental results showed we can achieve excellent classification ability even if no training samples are available. The source code is available at https://github.com/guizilaile23/ZSL-SSS .