Target Detection and Segmentation in Circular-Scan Synthetic-Aperture-Sonar Images using Semi-Supervised Convolutional Encoder-Decoders

TL;DR

This paper introduces a semi-supervised convolutional encoder-decoder framework for detecting and segmenting multiple targets in circular-scan synthetic-aperture-sonar images, outperforming existing natural-image-based methods.

Contribution

The paper presents a novel multi-branch encoder-decoder network tailored for CSAS imagery, integrating dual decoders and deep parsing for improved saliency-based target segmentation.

Findings

Outperforms supervised deep-saliency networks on CSAS data

Significantly better than unsupervised saliency methods for sonar images

Natural-image models require adaptation for sonar imagery

Abstract

We propose a framework for saliency-based, multi-target detection and segmentation of circular-scan, synthetic-aperture-sonar (CSAS) imagery. Our framework relies on a multi-branch, convolutional encoder-decoder network (MB-CEDN). The encoder portion of the MB-CEDN extracts visual contrast features from CSAS images. These features are fed into dual decoders that perform pixel-level segmentation to mask targets. Each decoder provides different perspectives as to what constitutes a salient target. These opinions are aggregated and cascaded into a deep-parsing network to refine the segmentation. We evaluate our framework using real-world CSAS imagery consisting of five broad target classes. We compare against existing approaches from the computer-vision literature. We show that our framework outperforms supervised, deep-saliency networks designed for natural imagery. It greatly outperforms unsupervised saliency approaches developed for natural imagery. This illustrates that natural-image-based models may need to be altered to be effective for this imaging-sonar modality.