RUIG: Realistic Underwater Image Generation Towards Restoration

In this paper, we present a novel method for generating synthetic underwater images considering revised image formation model. We propose to use the generated synthetic underwater images to train a conditional generative adversarial network (CGAN) towards restoration of degraded underwater images. Restoration of degraded underwater images using traditional dehazing models is challenging as they are insensitive to wavelength, depth, water type and treat backscattering and direct signal attenuation coefficients to be equal. However, learning based models for restoration perform well but sensitive to availability of ground truth information. Generating ground truth labels in underwater scenario demands in-situ measurements using expensive equipments and is infeasible due to varying underwater currents. Towards this, we propose to generate synthetic underwater images using revised image formation model. Revised image formation model is sensitive to different attenuation coefficients: 1) back scattering, 2) direct scattering and 3) veiling light. We propose to estimate these attenuation coefficients considering proven facts from the literature. We demonstrate restoration of real underwater images through restoration framework trained using rendered synthetic underwater images, and compare results of restoration with state-of-the-art techniques.

1. Introduction

In this paper, we propose a framework for generation of synthetic underwater images considering revised image formation model [1] and use the same to train conditional generative adversarial networks, towards restoration of degraded underwater images. Capturing of underwater scene, heavily relies on unmanned vehicles (UV) equipped with imaging sensors, to provide a high-resolution view of sea bed, corals and archaeological sites. Marine archaeologists use the remotely operated vehicle (ROV) to explore the ocean without physically being present in the ocean [12]. Recently, we observe considerable advancement in underwater scene capturing technologies. However, the aquatic environment still presents unique challenges, unlike the above-water environment. Due to light attenuation, absorption and scattering most of the underwater images lack in contrast and depict inaccurate colors. The attenuation of light in water varies with wavelength and depends on its distance, unlike the terrestrial images where attenuation is assumed to be spectrally uniform. Wavelength-dependent attenuation causes color distortion that increases with the distance of an object from the camera. This phenomenon causes underwater images to appear bluish or greenish in color, unlike the above water scene.