Towards Practical and Efficient High-Resolution HDR Deghosting with CNN

Prabhakar, KR and Agrawal, S and Singh, DK and Ashwath, B and Babu, RV (2020) Towards Practical and Efficient High-Resolution HDR Deghosting with CNN. In: Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), 23 - 28 August 2020, Glasgow, pp. 497-513.

PDF ECCV_2020.pdf - Published Version Restricted to Registered users only Download (9MB) | Request a copy

Abstract

Generating High Dynamic Range (HDR) image in the presence of camera and object motion is a tedious task. If uncorrected, these motions will manifest as ghosting artifacts in the fused HDR image. On one end of the spectrum, there exist methods that generate high-quality results that are computationally demanding and too slow. On the other end, there are few faster methods that produce unsatisfactory results. With ever increasing sensor/display resolution, currently we are very much in need of faster methods that produce high-quality images. In this paper, we present a deep neural network based approach to generate high-quality ghost-free HDR for high-resolution images. Our proposed method is fast and fuses a sequence of three high-resolution images (16-megapixel resolution) in about 10 s. Through experiments and ablations, on different publicly available datasets, we show that the proposed method achieves state-of-the-art performance in terms of accuracy and speed. © 2020, Springer Nature Switzerland AG.

Item Type:	Conference Paper
Publication:	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Publisher:	Springer Science and Business Media Deutschland GmbH
Additional Information:	The copyright for this article belongs to Springer Science and Business Media Deutschland GmbH.
Keywords:	Deep neural networks, Ghosting artifacts; High dynamic range images; High quality images; High resolution; High resolution image; Network-based approach; Object motion; State-of-the-art performance, Computer vision
Department/Centre:	Division of Interdisciplinary Sciences > Computational and Data Sciences
Date Deposited:	07 Feb 2023 09:18
Last Modified:	07 Feb 2023 09:18
URI:	https://eprints.iisc.ac.in/id/eprint/79994

Actions (login required)

View Item