SLYKLatent: A Learning Framework for Gaze Estimation Using Deep Facial Feature Learning
Fuente:
arXiv
Salvato in:
| Autori principali: | , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2024
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866912687977398272 |
|---|---|
| author | Adebayo, Samuel Dessing, Joost C. McLoone, Seán |
| author_facet | Adebayo, Samuel Dessing, Joost C. McLoone, Seán |
| contents | In this research, we present SLYKLatent, a novel approach for enhancing gaze estimation by addressing appearance instability challenges in datasets due to aleatoric uncertainties, covariant shifts, and test domain generalization. SLYKLatent utilizes Self-Supervised Learning for initial training with facial expression datasets, followed by refinement with a patch-based tri-branch network and an inverse explained variance-weighted training loss function. Our evaluation on benchmark datasets achieves a 10.9% improvement on Gaze360, supersedes top MPIIFaceGaze results with 3.8%, and leads on a subset of ETH-XGaze by 11.6%, surpassing existing methods by significant margins. Adaptability tests on RAF-DB and Affectnet show 86.4% and 60.9% accuracies, respectively. Ablation studies confirm the effectiveness of SLYKLatent's novel components. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2402_01555 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | SLYKLatent: A Learning Framework for Gaze Estimation Using Deep Facial Feature Learning Adebayo, Samuel Dessing, Joost C. McLoone, Seán Computer Vision and Pattern Recognition Artificial Intelligence Human-Computer Interaction Image and Video Processing In this research, we present SLYKLatent, a novel approach for enhancing gaze estimation by addressing appearance instability challenges in datasets due to aleatoric uncertainties, covariant shifts, and test domain generalization. SLYKLatent utilizes Self-Supervised Learning for initial training with facial expression datasets, followed by refinement with a patch-based tri-branch network and an inverse explained variance-weighted training loss function. Our evaluation on benchmark datasets achieves a 10.9% improvement on Gaze360, supersedes top MPIIFaceGaze results with 3.8%, and leads on a subset of ETH-XGaze by 11.6%, surpassing existing methods by significant margins. Adaptability tests on RAF-DB and Affectnet show 86.4% and 60.9% accuracies, respectively. Ablation studies confirm the effectiveness of SLYKLatent's novel components. |
| title | SLYKLatent: A Learning Framework for Gaze Estimation Using Deep Facial Feature Learning |
| topic | Computer Vision and Pattern Recognition Artificial Intelligence Human-Computer Interaction Image and Video Processing |
| url | https://arxiv.org/abs/2402.01555 |