Low-cost, vision-centric 3D perception systems for autonomous driving have made significant progress in recent years, narrowing the gap to expensive LiDAR-based methods. The primary challenge in becoming a fully reliable alternative lies in robust depth prediction capabilities, as camera-based systems struggle with long detection ranges and adverse lighting and weather conditions. In this work, we introduce HyDRa, a novel camera-radar fusion architecture for diverse 3D perception tasks. Building upon the principles of dense Bird's-EyeView (BEV)-based architectures, HyDRa introduces a hybrid fusion approach to combine the strengths of complementary camera and radar features in two distinct representation spaces. Our Height Association Transformer module leverages radar features already in the perspective view to produce more robust and accurate depth predictions. In the BEV, we refine the initial sparse representation by a Radar-weighted Depth Consistency. HyDRa achieves a new state-of-the-art for cameraradar fusion of 64.2 NDS (+1.8) and 58.4 AMOTA (+1.5) on the public nuScenes dataset. Moreover, our new semantically rich and spatially accurate BEV features can be directly converted into a powerful occupancy representation, beating all previous camera-based methods on the Occ3D benchmark by an impressive 3.7 mIoU. Code and models are available at https://github.com/phi-wol/hydra.

Object detectors are at the heart of many semi- and fully autonomous decision systems and are poised to become even more indispensable. They are, however, still lacking in accessibility and can sometimes produce unreliable predictions. Especially concerning in this regard are the—essentially hand-crafted—non-maximum suppression algorithms that lead to an obfuscated prediction process and biased confidence estimates. We show that we can eliminate classic NMS-style post-processing by using IoU-aware calibration. IoU-aware calibration is a conditional Beta calibration; this makes it parallelizable with no hyper-parameters. Instead of arbitrary cutoffs or discounts, it implicitly accounts for the likelihood of each detection being a duplicate and adjusts the confidence score accordingly, resulting in empirically based precision estimates for each detection. Our extensive experiments on diverse detection architectures show that the proposed IoU-aware calibration can successfully model duplicate detections and improve calibration. Compared to the standard sequential NMS and calibration approach, our joint modeling can deliver performance gains over the best NMS-based alternative while producing consistently better-calibrated confidence predictions with less complexity. The code for all our experiments is publicly available1.

Event cameras generate data based on the amount of motion present in the captured scene, making them attractive sensors for solving object tracking tasks. In this paper, we present a framework for tracking humans using a single event camera which consists of three components. First, we train a Graph Neural Network (GNN) to recognize a person within the stream of events. Batches of events are represented as spatio-temporal graphs in order to preserve the sparse nature of events and retain their high temporal resolution. Subsequently, the person is localized in a weakly-supervised manner by adopting the well established method of Class Activation Maps (CAM) for our graph-based classification model. Our approach does not require the ground truth position of humans during training. Finally, a Kalman filter is deployed for tracking, which uses the predicted bounding box surrounding the human as measurement. We demonstrate that our approach achieves robust tracking results on test sequences from the Gait3 database, paving the way for further privacy-preserving methods in event-based human tracking. Code, pre-trained models and datasets of our research are publicly available1.

Countless applications depend on accurate predictions with reliable confidence estimates from modern object detectors. However, it is well known that neural networks, including object detectors, produce miscalibrated confidence estimates. Recent work even suggests that detectors’ confidence predictions are biased with respect to object size and position. In object detection, the issues of conditional biases, confidence calibration, and task performance are usually explored in isolation, but, as we aim to show, they are closely related. We formally prove that the conditional confidence bias harms the performance of object detectors and empirically validate these findings. Specifically, to quantify the performance impact of the confidence bias on object detectors, we modify the histogram binning calibration to avoid performance impairment and instead improve it through calibration conditioned on the bounding box size. We further find that the confidence bias is also present in detections generated on the training data of the detector, which can be leveraged to perform the de-biasing. Moreover, we show that Test Time Augmentation (TTA) confounds this bias, which results in even more significant performance impairments on the detectors. Finally, we use our proposed algorithm to analyze a diverse set of object detection architectures and show that the conditional confidence bias harms their performance by up to 0.6 mAP and 0.8 mAP50. Code available at https://github.com/Blueblue4/Object-Detection-Confidence-Bias.

Person Re-Identification aims to retrieve person identities from images captured by multiple cameras or the same cameras in different time instances and locations. Because of its importance in many vision applications from surveillance to human-machine interaction, person re-identification methods need to be reliable and fast. While more and more deep architectures are proposed for increasing performance, those methods also increase overall model complexity. This paper proposes a lightweight network that combines global, part-based, and channel features in a unified multi-branch architecture that builds on the resource-efficient OSNet backbone. Using a well-founded combination of training techniques and design choices, our final model achieves state-of-the-art results on CUHK03 labeled, CUHK03 detected, and Market-1501 with 85.1% mAP/ 87.2% rankl, 82.4% mAP/84.9% rankl, and 91.5% mAP/96.3% rankl, respectively.

Typical approaches for video face recognition aggregate faces in a feature space to obtain a single feature representing the entire video. Unlike most previous approaches, we aggregate the faces directly in order to additionally obtain a single representative face as an intermediate output, from which a more discriminative feature vector is extracted. To overcome the limitation of a fixed number of input images of the state of the art in face aggregation, we incorporate a permutation invariant U-Net architecture capable of processing an arbitrary number of frames, which is employed in a generative adversarial network. We demonstrate the effectiveness of our method on three popular benchmark datasets for video face recognition. Our approach outperforms the baselines on the YouTube Faces dataset, obtaining an accuracy of 96.62%. Besides, we show that our method is robust against motion blur.

Gait recognition is a promising video-based biometric for identifying individual walking patterns from a long distance. At present, most gait recognition methods use silhouette images to represent a person in each frame. However, silhouette images can lose fine-grained spatial information, and most papers do not regard how to obtain these silhouettes in complex scenes. Furthermore, silhouette images contain not only gait features but also other visual clues that can be recognized. Hence these approaches can not be considered as strict gait recognition. We leverage recent advances in human pose estimation to estimate robust skeleton poses directly from RGB images to bring back model-based gait recognition with a cleaner representation of gait. Thus, we propose GaitGraph that combines skeleton poses with Graph Convolutional Network (GCN) to obtain a modern model-based approach for gait recognition. The main advantages are a cleaner, more elegant extraction of the gait features and the ability to incorporate powerful spatiotemporal modeling using GCN. Experiments on the popular CASIA-B gait dataset show that our method archives state-of-the-art performance in model-based gait recognition.The code and models are publicly available1