Training Matting Models without Alpha Labels
| Title: | Training Matting Models without Alpha Labels |
|---|---|
| Authors: | Liu, Wenze, Ye, Zixuan, Lu, Hao, Cao, Zhiguo, Yue, Xiangyu |
| Publication Year: | 2024 |
| Collection: | Computer Science |
| Subject Terms: | Computer Science - Computer Vision and Pattern Recognition |
| More Details: | The labelling difficulty has been a longstanding problem in deep image matting. To escape from fine labels, this work explores using rough annotations such as trimaps coarsely indicating the foreground/background as supervision. We present that the cooperation between learned semantics from indicated known regions and proper assumed matting rules can help infer alpha values at transition areas. Inspired by the nonlocal principle in traditional image matting, we build a directional distance consistency loss (DDC loss) at each pixel neighborhood to constrain the alpha values conditioned on the input image. DDC loss forces the distance of similar pairs on the alpha matte and on its corresponding image to be consistent. In this way, the alpha values can be propagated from learned known regions to unknown transition areas. With only images and trimaps, a matting model can be trained under the supervision of a known loss and the proposed DDC loss. Experiments on AM-2K and P3M-10K dataset show that our paradigm achieves comparable performance with the fine-label-supervised baseline, while sometimes offers even more satisfying results than human-labelled ground truth. Code is available at \url{https://github.com/poppuppy/alpha-free-matting}. Comment: 12 pages, 12 figures |
| Document Type: | Working Paper |
| Access URL: | http://arxiv.org/abs/2408.10539 |
| Accession Number: | edsarx.2408.10539 |
| Database: | arXiv |
| FullText | Text: Availability: 0 CustomLinks: – Url: http://arxiv.org/abs/2408.10539 Name: EDS - Arxiv Category: fullText Text: View this record from Arxiv MouseOverText: View this record from Arxiv – Url: https://resolver.ebsco.com/c/xy5jbn/result?sid=EBSCO:edsarx&genre=article&issn=&ISBN=&volume=&issue=&date=20240820&spage=&pages=&title=Training Matting Models without Alpha Labels&atitle=Training%20Matting%20Models%20without%20Alpha%20Labels&aulast=Liu%2C%20Wenze&id=DOI: Name: Full Text Finder (for New FTF UI) (s8985755) Category: fullText Text: Find It @ SCU Libraries MouseOverText: Find It @ SCU Libraries |
|---|---|
| Header | DbId: edsarx DbLabel: arXiv An: edsarx.2408.10539 RelevancyScore: 1112 AccessLevel: 3 PubType: Report PubTypeId: report PreciseRelevancyScore: 1112.24450683594 |
| IllustrationInfo | |
| Items | – Name: Title Label: Title Group: Ti Data: Training Matting Models without Alpha Labels – Name: Author Label: Authors Group: Au Data: <searchLink fieldCode="AR" term="%22Liu%2C+Wenze%22">Liu, Wenze</searchLink><br /><searchLink fieldCode="AR" term="%22Ye%2C+Zixuan%22">Ye, Zixuan</searchLink><br /><searchLink fieldCode="AR" term="%22Lu%2C+Hao%22">Lu, Hao</searchLink><br /><searchLink fieldCode="AR" term="%22Cao%2C+Zhiguo%22">Cao, Zhiguo</searchLink><br /><searchLink fieldCode="AR" term="%22Yue%2C+Xiangyu%22">Yue, Xiangyu</searchLink> – Name: DatePubCY Label: Publication Year Group: Date Data: 2024 – Name: Subset Label: Collection Group: HoldingsInfo Data: Computer Science – Name: Subject Label: Subject Terms Group: Su Data: <searchLink fieldCode="DE" term="%22Computer+Science+-+Computer+Vision+and+Pattern+Recognition%22">Computer Science - Computer Vision and Pattern Recognition</searchLink> – Name: Abstract Label: Description Group: Ab Data: The labelling difficulty has been a longstanding problem in deep image matting. To escape from fine labels, this work explores using rough annotations such as trimaps coarsely indicating the foreground/background as supervision. We present that the cooperation between learned semantics from indicated known regions and proper assumed matting rules can help infer alpha values at transition areas. Inspired by the nonlocal principle in traditional image matting, we build a directional distance consistency loss (DDC loss) at each pixel neighborhood to constrain the alpha values conditioned on the input image. DDC loss forces the distance of similar pairs on the alpha matte and on its corresponding image to be consistent. In this way, the alpha values can be propagated from learned known regions to unknown transition areas. With only images and trimaps, a matting model can be trained under the supervision of a known loss and the proposed DDC loss. Experiments on AM-2K and P3M-10K dataset show that our paradigm achieves comparable performance with the fine-label-supervised baseline, while sometimes offers even more satisfying results than human-labelled ground truth. Code is available at \url{https://github.com/poppuppy/alpha-free-matting}.<br />Comment: 12 pages, 12 figures – Name: TypeDocument Label: Document Type Group: TypDoc Data: Working Paper – Name: URL Label: Access URL Group: URL Data: <link linkTarget="URL" linkTerm="http://arxiv.org/abs/2408.10539" linkWindow="_blank">http://arxiv.org/abs/2408.10539</link> – Name: AN Label: Accession Number Group: ID Data: edsarx.2408.10539 |
| PLink | https://login.libproxy.scu.edu/login?url=https://search.ebscohost.com/login.aspx?direct=true&site=eds-live&scope=site&db=edsarx&AN=edsarx.2408.10539 |
| RecordInfo | BibRecord: BibEntity: Subjects: – SubjectFull: Computer Science - Computer Vision and Pattern Recognition Type: general Titles: – TitleFull: Training Matting Models without Alpha Labels Type: main BibRelationships: HasContributorRelationships: – PersonEntity: Name: NameFull: Liu, Wenze – PersonEntity: Name: NameFull: Ye, Zixuan – PersonEntity: Name: NameFull: Lu, Hao – PersonEntity: Name: NameFull: Cao, Zhiguo – PersonEntity: Name: NameFull: Yue, Xiangyu IsPartOfRelationships: – BibEntity: Dates: – D: 20 M: 08 Type: published Y: 2024 |
| ResultId | 1 |