TaskGuard predicts when image fixes help object detectors in poor conditions
TaskGuard: Task-Conditioned Restoration Utility for Risk-Aware Object Detection
Computer Vision and Pattern Recognition
Summary
Object detection systems often use image cleanup methods to help recognize objects in blurry or rainy pictures, but these fixes don’t always improve results. The authors propose TaskGuard, a tool that decides whether to keep the original image or use the fixed version for detection. It learns from how different fix attempts affect detection performance and works well even on conditions it wasn’t specifically trained on. This helps avoid using image corrections that actually make object detection worse.
What this means in practice
- •For autonomous vehicle teams: Decide when to apply image restoration before object detection to maintain better detection accuracy in bad weather or motion conditions.
- •For surveillance system engineers: Improve reliability by selectively using restored images only when they help object detectors in scenes with rain, blur, or defocus.
Authors
Vung Pham
Abstract
Image restoration is commonly applied before object detection under adverse conditions, yet a visually improved image need not improve the downstream task. We study this mismatch as restoration utility prediction: given a degraded image and its candidate restoration, should the restoration be used or should the original observation be preserved? We introduce TaskGuard, a post-hoc controller for frozen restoration and detection pipelines. TaskGuard characterizes the realized restoration residual through its interaction with detector sensitivity and predicts whether the intervention is task-beneficial. Exact regional counterfactuals reveal substantial within-image utility heterogeneity, while a deployable pseudo-gradient preserves statistically reliable directional information. Feature-group ablation further shows that task-conditioned evidence contributes information beyond detector-response and residual statistics. The TaskGuard utility predictor is trained only on Gaussian degradation and frozen before final evaluation, then transferred to unseen motion blur, rain, and defocus. Across these unseen families, TaskGuard reduces lossnegative interventions by 54.2% (family macro) and practical per-image detection deteriorations by 37.0% (pooled), while preserving 98.8% of the Always-Restore COCO AP. On natural-rain DAWN, it reduces loss-negative interventions by 97.9% while retaining 77.8% of the AP improvement obtained by deraining. These results support restoration utility as a task-conditioned property of the specific intervention rather than image appearance alone.