Towards Adapting CLIP for Gaze Object Prediction
Proceedings of the ACM on Human-Computer Interaction · 24.10.2024 · 8 (ISS) · 155-164
DOI: 10.1145/3698132 ↗Kurz gefasst
Fachartikel in der Zeitschrift Proceedings of the ACM on Human-Computer Interaction, online erschienen am 24.10.2024. Verfasst von Dazhi Chen und 1 weiteren Autor:innen. Die Arbeit wurde bisher 1-mal zitiert (Crossref) und nennt 22 Literaturangaben.
Abstract (Originalsprache)
This paper aims to investigate the problem of gaze object prediction in single images. We propose an application-friendly network based on CLIP for gaze object prediction. To avoid domain bias, we utilize a shallow feature adapter that transfers pre-trained features to target-oriented ones. Secondly, we introduce a pooling attention block to exploit the joint representation of multimodal elements, reducing gaze point deviation. Additionally, we introduce a loss that measures the prediction quality by comparing the distribution difference between the model's predictions heatmaps and the ground truth. Extensive experiments demonstrate the superior performance of our model compared to previous models. We will provide the method code at: https://github.com/fadaishaitaiyang/CCLIP.git.
Zitieren
APA
Chen, D., & Gou, G. (2024). Towards Adapting CLIP for Gaze Object Prediction. Proceedings of the ACM on Human-Computer Interaction, 8(ISS), 155-164. https://doi.org/10.1145/3698132
Vancouver
Chen D, Gou G. Towards Adapting CLIP for Gaze Object Prediction. Proceedings of the ACM on Human-Computer Interaction. 2024;8(ISS):155-164. doi:10.1145/3698132
BibTeX
@article{Chen2024,
author = {Chen, Dazhi and Gou, Gang},
title = {Towards Adapting CLIP for Gaze Object Prediction},
journal = {Proceedings of the ACM on Human-Computer Interaction},
year = {2024},
volume = {8},
number = {ISS},
pages = {155--164},
doi = {10.1145/3698132}
}
EUVIMED