Reflection Removal under Fast Forward Camera Motion

Jun Young Cheong, Christian Simon, Chang-Su Kim, In Kyu Park

Research output: Contribution to journalArticle

1 Citation (Scopus)

Abstract

The image quality of an in-vehicle black box camera is often degraded by reflections of internal objects, dirt, and dust on the windshield. In this paper, we propose a novel algorithm that simultaneously removes the reflections and small dirt artifacts from in-vehicle black box videos under fast forward camera motion. The algorithm exploits the spatiotemporal coherence of the reflection and dirt, which remain stationary relative to the fast-moving background. Unlike previous algorithms, the algorithm first separates stationary reflection and then restores the background scene. To this end, we propose an average image prior, thereby imposing spatiotemporal coherence. The separation model is a two-layer model composed of stationary and background layers, where different gradient sparsity distributions are utilized in a region-based manner. Motion compensation in postprocessing is proposed to alleviate layer jitter due to vehicle vibrations. In evaluation experiments, the proposed algorithm successfully extracts the stationary layer from several real and synthetic black box videos.

Original languageEnglish
JournalIEEE Transactions on Image Processing
DOIs
Publication statusAccepted/In press - 2017 Aug 31

Fingerprint

Cameras
Windshields
Motion compensation
Vibration
Dust
Jitter
Artifacts
Image quality
Experiments

Keywords

  • Automobiles
  • Automotive components
  • average image prior
  • Black box camera
  • Cameras
  • dirt removal
  • fast forward motion
  • Glass
  • Image reconstruction
  • layer separation
  • Optical imaging
  • reflection removal
  • Videos

ASJC Scopus subject areas

  • Software
  • Medicine(all)
  • Computer Graphics and Computer-Aided Design

Cite this

Reflection Removal under Fast Forward Camera Motion. / Cheong, Jun Young; Simon, Christian; Kim, Chang-Su; Park, In Kyu.

In: IEEE Transactions on Image Processing, 31.08.2017.

Research output: Contribution to journalArticle

@article{f85a2db7b7a9447c999802c8be639101,
title = "Reflection Removal under Fast Forward Camera Motion",
abstract = "The image quality of an in-vehicle black box camera is often degraded by reflections of internal objects, dirt, and dust on the windshield. In this paper, we propose a novel algorithm that simultaneously removes the reflections and small dirt artifacts from in-vehicle black box videos under fast forward camera motion. The algorithm exploits the spatiotemporal coherence of the reflection and dirt, which remain stationary relative to the fast-moving background. Unlike previous algorithms, the algorithm first separates stationary reflection and then restores the background scene. To this end, we propose an average image prior, thereby imposing spatiotemporal coherence. The separation model is a two-layer model composed of stationary and background layers, where different gradient sparsity distributions are utilized in a region-based manner. Motion compensation in postprocessing is proposed to alleviate layer jitter due to vehicle vibrations. In evaluation experiments, the proposed algorithm successfully extracts the stationary layer from several real and synthetic black box videos.",
keywords = "Automobiles, Automotive components, average image prior, Black box camera, Cameras, dirt removal, fast forward motion, Glass, Image reconstruction, layer separation, Optical imaging, reflection removal, Videos",
author = "Cheong, {Jun Young} and Christian Simon and Chang-Su Kim and Park, {In Kyu}",
year = "2017",
month = "8",
day = "31",
doi = "10.1109/TIP.2017.2748389",
language = "English",
journal = "IEEE Transactions on Image Processing",
issn = "1057-7149",
publisher = "Institute of Electrical and Electronics Engineers Inc.",

}

TY - JOUR

T1 - Reflection Removal under Fast Forward Camera Motion

AU - Cheong, Jun Young

AU - Simon, Christian

AU - Kim, Chang-Su

AU - Park, In Kyu

PY - 2017/8/31

Y1 - 2017/8/31

N2 - The image quality of an in-vehicle black box camera is often degraded by reflections of internal objects, dirt, and dust on the windshield. In this paper, we propose a novel algorithm that simultaneously removes the reflections and small dirt artifacts from in-vehicle black box videos under fast forward camera motion. The algorithm exploits the spatiotemporal coherence of the reflection and dirt, which remain stationary relative to the fast-moving background. Unlike previous algorithms, the algorithm first separates stationary reflection and then restores the background scene. To this end, we propose an average image prior, thereby imposing spatiotemporal coherence. The separation model is a two-layer model composed of stationary and background layers, where different gradient sparsity distributions are utilized in a region-based manner. Motion compensation in postprocessing is proposed to alleviate layer jitter due to vehicle vibrations. In evaluation experiments, the proposed algorithm successfully extracts the stationary layer from several real and synthetic black box videos.

AB - The image quality of an in-vehicle black box camera is often degraded by reflections of internal objects, dirt, and dust on the windshield. In this paper, we propose a novel algorithm that simultaneously removes the reflections and small dirt artifacts from in-vehicle black box videos under fast forward camera motion. The algorithm exploits the spatiotemporal coherence of the reflection and dirt, which remain stationary relative to the fast-moving background. Unlike previous algorithms, the algorithm first separates stationary reflection and then restores the background scene. To this end, we propose an average image prior, thereby imposing spatiotemporal coherence. The separation model is a two-layer model composed of stationary and background layers, where different gradient sparsity distributions are utilized in a region-based manner. Motion compensation in postprocessing is proposed to alleviate layer jitter due to vehicle vibrations. In evaluation experiments, the proposed algorithm successfully extracts the stationary layer from several real and synthetic black box videos.

KW - Automobiles

KW - Automotive components

KW - average image prior

KW - Black box camera

KW - Cameras

KW - dirt removal

KW - fast forward motion

KW - Glass

KW - Image reconstruction

KW - layer separation

KW - Optical imaging

KW - reflection removal

KW - Videos

UR - http://www.scopus.com/inward/record.url?scp=85029009341&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85029009341&partnerID=8YFLogxK

U2 - 10.1109/TIP.2017.2748389

DO - 10.1109/TIP.2017.2748389

M3 - Article

JO - IEEE Transactions on Image Processing

JF - IEEE Transactions on Image Processing

SN - 1057-7149

ER -