Detection Transformer (DETR) and its variants show strong performance on object detection, a key task for autonomous systems. However, a critical limitation of these models is that their confidence scores only reflect semantic uncertainty, failing to capture the equally important spatial uncertainty. This results in an incomplete assessment of the detection reliability. On the other hand, Deep Ensembles can tackle this by providing high-quality spatial uncertainty estimates. However, their immense memory consumption makes them impractical for real-world applications. A cheaper alternative, Monte Carlo (MC) Dropout, suffers from high latency due to the need of multiple forward passes during inference to estimate uncertainty. To address these limitations, we introduce GroupEnsemble, an efficient and effective uncertainty estimation method for DETR-like models. GroupEnsemble simultaneously predicts multiple individual detection sets by feeding additional diverse groups of object queries to the transformer decoder during inference. Each query group is transformed by the shared decoder in isolation and predicts a complete detection set for the same input. An attention mask is applied to the decoder to prevent inter-group query interactions, ensuring each group detects independently to achieve reliable ensemble-based uncertainty estimation. By leveraging the decoder's inherent parallelism, GroupEnsemble efficiently estimates uncertainty in a single forward pass without sequential repetition. We validated our method under autonomous driving scenes and common daily scenes using the Cityscapes and COCO datasets, respectively. The results show that a hybrid approach combining MC-Dropout and GroupEnsemble outperforms Deep Ensembles on several metrics at a fraction of the cost. The code is available at https://github.com/yutongy98/GroupEnsemble.
Figures & tables
Fig. 1: Visualizations of semantic and spatial uncertainties estimated by GroupEnsemble . Semantic uncertainty is represented by the classification confidence (Class. Conf.), while spatial uncertainty is visualized using 95% confidence intervals (dashed lines) around the mean (solid line). The color coding indicates the level of uncertainty: green for certain detections, yellow for medium uncertainty, and red for high uncertainty. Factors such as occlusion or background clutter can lead to uncertainty in the location or size of the detected objects. By estimating both uncertainties, GroupEnsemble provides a comprehensive assessment of the reliability of the detections.
Fig. 2: Overview of GroupEnsemble. For an input image, a standard DETR (a) outputs deterministic detections. In contrast, our GroupEnsemble (b) feeds G−1 additional groups of object queries to the decoder and employs a self-attention mask (gray grids indicate blocked attentions) to block all inter-group query interactions. This enables the decoder to independently and simultaneously transform each query group, predicting multiple individual detection sets in a single pass. These detections are then clustered and aggregated to produce final detections with semantic and spatial uncertainty estimates. Spatial uncertainty is depicted as dashed confidence intervals around the mean box. Larger intervals indicate higher uncertainty. e.g., for the person occluded by the truck door.
Fig. 3: The detection performance per query group is similar. The mAP scores are obtained by training a Conditional DETR [ 15 ] with Group DETR [ 18 ] on the Cityscapes dataset [ 12 ] , using five query groups.
Fig. 4: Visualizations of detection clusters and the corresponding object queries (represented as center reference points) from different query groups on the Cityscapes dataset [ 12 ] . Different colors are used to distinguish between groups. The diversity of reference points detecting the same ground-truth objects is evident. The reference points and bounding boxes may overlap.
Dataset
Method
Uncertainty & Acc.
Calibration
Accuracy
Efficiency
↑ PDQ(thr. 0.3)
↓ D-ECE(thr. 0.3)
↑ mAP
↓ Latency(ms) ∗
↓ #Para(M)
Cityscapes [ 12 ]
Deterministic
9.4
11.8
37.8
10.7
43.17
MC-Dropout †
18.9
10.9
38.5
18.4( ×1.7 )
43.17
Deep Ensembles
19.3
9.8
38.8
53.4( \times{\color[rgb]{1,0,0}5.0} )
89.27 (+ 107 %)
GroupEnsemble (ours)
18.8
10.4
38.7
18.4( ×1.7 )
43.48 (+0.7%)
MC-GroupEnsemble (ours)
21.4
10.4
39.2
18.4( ×1.7 )
43.48 (+0.7%)
TABLE I: Comparison of GroupEnsemble with existing methods. ∗ measurements are reported for the transformer decoder only. † denotes our parallelized version. Best results in bold , second best underlined , highest latency and parameters in red.
Fig. 5: D-ECE plots of the different confidence aggregation methods. The calibration plots reveal that using the mean confidence score of clusters as the final scores results in under-confident detections (as highlighted in the 2nd plot), whereas using the maximum confidence without scaling leads to slightly over-confident detections (highlighted in the 3rd plot). Both cases negatively impact calibration performance. On the other hand, using the scaled maximum confidence, as shown in the 4th plot, effectively addresses these issues and improves calibration.
#Groups
↑ PDQ
↓ D-ECE
↑ mAP
↓ Latency(ms) ∗
↓ #Para(M)
1
9.4
11.8
37.8
10.68
43.17
3
17.5
10.6
38.3
15.42 (×1.4)
43.32
5
18.8
10.4
38.7
18.39 (×1.7)
43.48
7
19.4
10.4
38.7
22.51 (×2.1)
43.61
9
19.7
10.3
38.8
26.55 (×2.5)
43.76
TABLE II: Ablation study on the number of query groups for GroupEnsemble. ∗ Measurements are reported for the transformer decoder only. Best results in bold .
Conf. Aggregation
↑ PDQ
↓ D-ECE
↑ mAP
Baseline
9.4
11.8
37.8
Mean Conf.
16.2
14.1
37.6
Max Conf.
15.9
13.4
38.4
Max Conf. w/ Scaling
18.8
10.4
38.7
TABLE III: Ablation study on the aggregation methods for confidence scores. Best results in bold .
Object detection is a safety-critical component of autonomous driving. It is essential to quantify the uncertainty in bounding-box predictions for safety assurance. Post hoc uncertainty quantification without retraining aligns with real-world deployment requirements; therefore, we employ the Laplace approximation. Because instance-level uncertainty is needed, linearized inference methods that require multiple backpropagations are not time-efficient, and sampling-based methods are not fully post hoc. We propose Monte-Carlo generalized linearized model (MC-GLM), which provides instance-level and approximately post hoc uncertainty quantification. The number of samples required in the Monte Carlo step is constant and independent of the number of output instances, so it can be parallelized. Experiments on the nuScenes dataset with the CenterPoint detector validate the effectiveness of our method, and the resulting uncertainties exhibit good quality.
Chongzhe Zhang, Zifan Zeng, Qunli Zhang +2
RAMS Lab, Huawei Heisenberg Research Center, Munich, Germany · Technische Universität Berlin, Berlin, Germany · Technische Universität München, Munich, Germany
Reliable uncertainty estimation is essential for deploying object detectors when distribution/covariate shift and adversarial attacks may occur. Existing approaches often require detector retraining, architectural modification, or repeated inference, which may be infeasible or incur significant overheads. We introduce a lightweight post-hoc evidential meta-model that learns when object localisations should be considered uncertain while keeping the base detector frozen. Our approach automatically identifies localisation-relevant features and uses saliency-guided modification to construct an increasingly challenging curriculum. Detection-level targets combine localisation error, modification level, and prediction instability to guide an evidential meta-model to estimate uncertainty for each predicted bounding box. Our approach requires no changes to the detector and preserves its original localisation outputs. Across adversarial attacks and evaluated strengths, GRACE improves TP-FP AUROC by 22% relative to the strongest comparator in some cases while maintaining in-distribution detection performance.
Charmaine Barker, Daniel Bethell, Simos Gerasimou
Department of Computer Science, University of York, UK · Department of Elect. Eng., and Computer Science and Eng., Cyprus University of Technology, Cyprus
Reliable uncertainty estimation is essential for deploying object detectors in autonomous systems operating in uncertain environments. Evidential Deep Learning (EDL) provides a principled framework for uncertainty-aware classification by representing network outputs as evidence and interpreting predictions through subjective logic. However, existing evidential object detectors typically combine evidential classification with regression uncertainty models that do not share the same theoretical foundation. In this work, we propose an evidential version of YOLOv8 in which both classification and bounding-box regression are formulated within a common evidential framework. Our approach exploits YOLOv8's distribution-based bounding-box representation, allowing the evidential formulation to be applied not only to classification but also to localisation. As a result, both tasks produce belief, uncertainty, and probability estimates that can be interpreted within the Dempster--Shafer framework. Experiments on KITTI, MUSES, and nuScenes show that the resulting detector remains broadly competitive with standard YOLOv8 in terms of detection accuracy while providing a localisation uncertainty that effectively discriminates between correct and erroneous detections. Moreover, this uncertainty becomes increasingly discriminative under domain shift.
Simon Barbarit-Gaboriau, Hind Laghmara, Rémi Boutteau +1
LITIS - STI, INSA Rouen Normandie · LITIS - STI · LITIS, LITIS - STI