TaskGuard improves object detection by deciding when image fixes help
TaskGuard: Task-Conditioned Restoration Utility for Risk-Aware Object Detection
Computer Vision and Pattern Recognition
Summary
Sometimes fixing a blurry or rainy photo doesn't help a computer find objects better. The authors created TaskGuard, a tool that looks at a fixed photo and decides if using the fix will actually help find objects or if it’s better to keep the original. TaskGuard learns how the changes in the image affect the detector's ability and can predict when the fix is useful. It works well even on new kinds of image problems like rain and motion blur without extra training. This shows that whether fixing an image helps depends on the specific detection task, not just how the picture looks.
image restorationobject detectiontask-conditioned predictionGaussian degradationmotion blurrain removaldefocus blurutility predictiondetector sensitivitypost-hoc controller
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.