Evolving task-agnostic features from task-specific priors for multi-task dense prediction
Publication Type
Journal Article
Publication Date
3-2026
Abstract
Previous multi-task dense prediction methods based on a decoder-centric paradigm have achieved promising results. They primarily rely on task-agnostic features extracted from a pre-trained backbone, overlooking the importance of task-specific prior knowledge. In this paper, we propose a novel approach that integrates task priors into task-agnostic features to facilitate the development of a more refined task-specific representativeness. Specifically, we introduce a CNN-based task-prior awareness module to capture comprehensive task-specific priors. As the cornerstone of feature interaction, task-agnostic features (e.g., pretrained ViT tokens) recurrently absorb the extracted task priors, evolving into multi-task-informed features. In the meantime, multi-task-informed features enhance task-specific features in favor of dense prediction. Furthermore, to address the inherent incompatibility between ViT tokens and CNN-based priors, we incorporate multi-scale collaborative feature filters within the decoder that adaptively generate low-pass and high-pass kernels to modulate their collaboration. Extensive experiments on the PASCAL-Context and NYUD-v2 datasets demonstrate that our approach achieves comparable or even better performance than the existing state-of-the-art methods.
Keywords
Dense Prediction, Multi-Task Learning
Discipline
Artificial Intelligence and Robotics | Software Engineering
Publication
IEEE Transactions on Multimedia
First Page
1
Last Page
13
ISSN
1520-9210
Identifier
10.1109/TMM.2026.3673564
Publisher
Institute of Electrical and Electronics Engineers
Citation
LI, Qi; LUO, Jiexin; DENG, Wenqi; YANG, Wenjie; CHEN, Fei; YU, Yuanlong; PAN, Jia; HE, Shengfeng; and LIU, Wenxi.
Evolving task-agnostic features from task-specific priors for multi-task dense prediction. (2026). IEEE Transactions on Multimedia. 1-13.
Available at: https://ink.library.smu.edu.sg/sis_research/11071
Additional URL
https://doi.org/10.1109/TMM.2026.3673564