DC Field | Value | Language |
---|---|---|
dc.contributor.author | Kang, Seoungyoon | ko |
dc.contributor.author | Lee, Minhyun | ko |
dc.contributor.author | Kim, Minjae | ko |
dc.contributor.author | Shim, Hyunjung | ko |
dc.date.accessioned | 2023-05-02T08:01:20Z | - |
dc.date.available | 2023-05-02T08:01:20Z | - |
dc.date.created | 2023-05-02 | - |
dc.date.issued | 2023 | - |
dc.identifier.citation | IEEE ACCESS, v.11, pp.26125 - 26135 | - |
dc.identifier.issn | 2169-3536 | - |
dc.identifier.uri | http://hdl.handle.net/10203/306421 | - |
dc.description.abstract | Facial landmark detection is an essential task in face-processing techniques. Traditional methods, however, require expensive pixel-level labels. Semi-supervised facial landmark detection has been explored as an alternative, but previous approaches only focus on training-oriented issues (e.g., noisy pseudo-labels in semi-supervised learning), neglecting task-oriented issues (i.e., the quantization error in landmark detection). We argue that semi-supervised landmark detectors should resolve the two technical issues simultaneously. Through a simple experiment, we found that task- and training-oriented solutions may negatively influence each other, thus eliminating their negative interactions is important. To this end, we devise a new heatmap regression framework via hybrid representation, namely HybridMatch. We utilize both 1-D and 2-D heatmap representations. Here, the 1-D and 2-D heatmaps help alleviate the task-oriented and training-oriented issues, respectively. To exploit the advantages of our hybrid representation, we introduce curriculum learning; relying more on the 2-D heatmap at the early training stage and gradually increasing the effects of the 1-D heatmap. By resolving the two issues simultaneously, we can capture more precise landmark points than existing methods with only a few annotated data. Extensive experiments show that HybridMatch achieves state-of-the-art performance on three benchmark datasets, especially showing 26.3% NME improvement over the existing method in the 300-W full set at 5% data ratio. Surprisingly, our method records a comparable performance, 5.04 (challenging set in the 300-W) to the fully-supervised facial landmark detector 5.03. The remarkable performance of HybridMatch shows its potential as a practical alternative to the fully-supervised model. | - |
dc.language | English | - |
dc.publisher | IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC | - |
dc.title | HybridMatch: Semi-Supervised Facial Landmark Detection via Hybrid Heatmap Representations | - |
dc.type | Article | - |
dc.identifier.wosid | 000966398000001 | - |
dc.identifier.scopusid | 2-s2.0-85151374509 | - |
dc.type.rims | ART | - |
dc.citation.volume | 11 | - |
dc.citation.beginningpage | 26125 | - |
dc.citation.endingpage | 26135 | - |
dc.citation.publicationname | IEEE ACCESS | - |
dc.identifier.doi | 10.1109/ACCESS.2023.3257180 | - |
dc.contributor.localauthor | Shim, Hyunjung | - |
dc.contributor.nonIdAuthor | Kang, Seoungyoon | - |
dc.contributor.nonIdAuthor | Lee, Minhyun | - |
dc.contributor.nonIdAuthor | Kim, Minjae | - |
dc.description.isOpenAccess | N | - |
dc.type.journalArticle | Article | - |
dc.subject.keywordAuthor | Facial landmark detection | - |
dc.subject.keywordAuthor | facial key-points | - |
dc.subject.keywordAuthor | landmark detection | - |
dc.subject.keywordAuthor | semi-supervised facial landmark detection | - |
dc.subject.keywordAuthor | heatmap-based landmark detection | - |
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.