TaskIR improves images and tasks by adapting to damage types
TaskIR: Task-Driven Image Restoration via Degradation Adaptation and Task Feedback
Computer Vision and Pattern Recognition
Summary
Images taken in real life often have different kinds of damage that make them hard to see or analyze. The authors created TaskIR, a method that first figures out what kind of damage an image has and fixes it accordingly. Then, it looks at how well the fixed image works for other tasks and refines it further based on feedback. This dual approach helps make images clearer and better for tasks like recognizing objects, even when the damage is different each time.
What this means in practice
- •For computer vision engineers: Improve image preprocessing models to better handle varied image defects for enhanced object detection accuracy.
- •For surveillance system developers: Boost visual clarity and task performance of security cameras by adapting restoration to different real-world degradation types.$Commercial implications: Enables more reliable security camera outputs by intelligently restoring images for better automated monitoring and threat detection.
Authors
Yanjie Tu, Qingsen Yan, Axi Niu, Wenxuan Cai, Tao Hu, Wei Dong, Haokui Zhang
Abstract
Task-driven image restoration aims to improve both image quality and downstream task performance. However, existing methods predominantly focus on single degradation type and struggle to handle the diverse degradations encountered in real-world scenarios. Different degradations impose distinct restoration demands, and insufficient restoration may leave residual degradations and artifacts that impair object boundaries and semantic cues, thereby compromising downstream task performance. To address these challenges, we propose TaskIR, a two-stage task-driven unified image restoration framework that integrates degradation-adaptive restoration with task feedback refinement. In Stage I, a Degradation Representation Module (DRM) extracts degradation representations, enabling a Degradation-Guided Transformer Block (DGTB) to dynamically modulate feature transformations for adaptive restoration. In Stage II, a Task-to-Restoration Feedback Generation module (TRFG) transforms heterogeneous task features into restoration feedback by modeling task-representation discrepancies associated with the current restoration. Subsequently, a Selective Task Feedback Refinement module (STFR) assesses feedback relevance and selectively refines intermediate restoration features to mitigate interference with well-restored content. Extensive experiments demonstrate that TaskIR achieves competitive restoration quality and downstream task performance across diverse degradations and tasks.