Generative multiview inpainting for object removal in large indoor spaces. (9th March 2021)
- Record Type:
- Journal Article
- Title:
- Generative multiview inpainting for object removal in large indoor spaces. (9th March 2021)
- Main Title:
- Generative multiview inpainting for object removal in large indoor spaces
- Authors:
- Kim, Joohyung
Hyeon, Janghun
Doh, Nakju - Abstract:
- As interest in image-based rendering increases, the need for multiview inpainting is emerging. Despite of rapid progresses in single-image inpainting based on deep learning approaches, they have no constraint in obtaining color consistency over multiple inpainted images. We target object removal in large-scale indoor spaces and propose a novel pipeline of multiview inpainting to achieve color consistency and boundary consistency in multiple images. The first step of the pipeline is to create color prior information on masks by coloring point clouds from multiple images and projecting the colored point clouds onto the image planes. Next, a generative inpainting network accepts a masked image, a color prior image, imperfect guideline, and two different masks as inputs and yields the refined guideline and inpainted image as outputs. The color prior and guideline input ensure color and boundary consistencies across multiple images. We validate our pipeline on real indoor data sets quantitatively using consistency distance and similarity distance, metrics we defined for comparing results of multiview inpainting and qualitatively.
- Is Part Of:
- International journal of advanced robotic systems. Volume 18:Number 2(2021)
- Journal:
- International journal of advanced robotic systems
- Issue:
- Volume 18:Number 2(2021)
- Issue Display:
- Volume 18, Issue 2 (2021)
- Year:
- 2021
- Volume:
- 18
- Issue:
- 2
- Issue Sort Value:
- 2021-0018-0002-0000
- Page Start:
- Page End:
- Publication Date:
- 2021-03-09
- Subjects:
- Multiview inpainting -- object removal -- generative adversarial network -- color consistency -- boundary consistency
Robotics -- Periodicals
Robotics
Periodicals
629.892 - Journal URLs:
- http://arx.sagepub.com/ ↗
http://search.epnet.com/direct.asp?db=bch&jid=13CR&scope=site ↗
http://www.intechweb.org/journal.php?id=3 ↗
http://www.uk.sagepub.com/home.nav ↗ - DOI:
- 10.1177/1729881421996544 ↗
- Languages:
- English
- ISSNs:
- 1729-8806
- Deposit Type:
- Legaldeposit
- View Content:
- Available online (eLD content is only available in our Reading Rooms) ↗
- Physical Locations:
- British Library DSC - BLDSS-3PM
British Library HMNTS - ELD Digital store - Ingest File:
- 15774.xml