A Deep Automotive Radar Detector using the RaDelft Dataset

The detection of multiple extended targets in complex environments using high-resolution automotive radar is considered. A data-driven approach is proposed where unlabeled synchronized lidar data is used as ground truth to train a neural network with only radar data as input. To this end, the novel, large-scale, real-life, and multi-sensor RaDelft dataset has been recorded using a demonstrator vehicle in different locations in the city of Delft. The dataset, as well as the documentation and example code, is publicly available for those researchers in the field of automotive radar or machine perception. The proposed data-driven detector is able to generate lidar-like point clouds using only radar data from a high-resolution system, which preserves the shape and size of extended targets. The results are compared against conventional CFAR detectors as well as variations of the method to emulate the available approaches in the literature, using the probability of detection, the probability of false alarm, and the Chamfer distance as performance metrics. Moreover, an ablation study was carried out to assess the impact of Doppler and temporal information on detection performance. The proposed method outperforms the different baselines in terms of Chamfer distance, achieving a reduction of 75% against conventional CFAR detectors and 10% against the modified state-of-the-art deep learning-based approaches.

Adapting Fine-Grained Cross-View Localization to Areas without Fine Ground Truth

Given a ground-level query image and a geo-referenced aerial image that covers the query's local surroundings, fine-grained cross-view localization aims to estimate the location of the ground camera inside the aerial image. Recent works have focused on developing advanced networks trained with accurate ground truth (GT) locations of ground images. However, the trained models always suffer a performance drop when applied to images in a new target area that differs from training. In most deployment scenarios, acquiring fine GT, i.e. accurate GT locations, for target-area images to re-train the network can be expensive and sometimes infeasible. In contrast, collecting images with noisy GT with errors of tens of meters is often easy. Motivated by this, our paper focuses on improving the performance of a trained model in a new target area by leveraging only the target-area images without fine GT. We propose a weakly supervised learning approach based on knowledge self-distillation. This approach uses predictions from a pre-trained model as pseudo GT to supervise a copy of itself. Our approach includes a mode-based pseudo GT generation for reducing uncertainty in pseudo GT and an outlier filtering method to remove unreliable pseudo GT. Our approach is validated using two recent state-of-the-art models on two benchmarks. The results demonstrate that it consistently and considerably boosts the localization accuracy in the target area.

On the Estimation of Image-matching Uncertainty in Visual Place Recognition

In Visual Place Recognition (VPR) the pose of a query image is estimated by comparing the image to a map of reference images with known reference poses. As is typical for image retrieval problems, a feature extractor maps the query and reference images to a feature space, where a nearest neighbor search is then performed. However, till recently little attention has been given to quantifying the confidence that a retrieved reference image is a correct match. Highly certain but incorrect retrieval can lead to catastrophic failure of VPR-based localization pipelines. This work compares for the first time the main approaches for estimating the image-matching uncertainty, including the traditional retrieval-based uncertainty estimation, more recent data-driven aleatoric uncertainty estimation, and the compute-intensive geometric verification. We further formulate a simple baseline method, ``SUE'', which unlike the other methods considers the freely-available poses of the reference images in the map. Our experiments reveal that a simple L2-distance between the query and reference descriptors is already a better estimate of image-matching uncertainty than current data-driven approaches. SUE outperforms the other efficient uncertainty estimation methods, and its uncertainty estimates complement the computationally expensive geometric verification approach. Future works for uncertainty estimation in VPR should consider the baselines discussed in this work.

See Further Than CFAR: a Data-Driven Radar Detector Trained by Lidar

In this paper, we address the limitations of traditional constant false alarm rate (CFAR) target detectors in automotive radars, particularly in complex urban environments with multiple objects that appear as extended targets. We propose a data-driven radar target detector exploiting a highly efficient 2D CNN backbone inspired by the computer vision domain. Our approach is distinguished by a unique cross sensor supervision pipeline, enabling it to learn exclusively from unlabeled synchronized radar and lidar data, thus eliminating the need for costly manual object annotations. Using a novel large-scale, real-life multi-sensor dataset recorded in various driving scenarios, we demonstrate that the proposed detector generates dense, lidar-like point clouds, achieving a lower Chamfer distance to the reference lidar point clouds than CFAR detectors. Overall, it significantly outperforms CFAR baselines detection accuracy.

MuVieCAST: Multi-View Consistent Artistic Style Transfer

We introduce MuVieCAST, a modular multi-view consistent style transfer network architecture that enables consistent style transfer between multiple viewpoints of the same scene. This network architecture supports both sparse and dense views, making it versatile enough to handle a wide range of multi-view image datasets. The approach consists of three modules that perform specific tasks related to style transfer, namely content preservation, image transformation, and multi-view consistency enforcement. We extensively evaluate our approach across multiple application domains including depth-map-based point cloud fusion, mesh reconstruction, and novel-view synthesis. Our experiments reveal that the proposed framework achieves an exceptional generation of stylized images, exhibiting consistent outcomes across perspectives. A user study focusing on novel-view synthesis further confirms these results, with approximately 68\% of cases participants expressing a preference for our generated outputs compared to the recent state-of-the-art method. Our modular framework is extensible and can easily be integrated with various backbone architectures, making it a flexible solution for multi-view style transfer. More results are demonstrated on our project page: muviecast.github.io

UniBEV: Multi-modal 3D Object Detection with Uniform BEV Encoders for Robustness against Missing Sensor Modalities

Multi-sensor object detection is an active research topic in automated driving, but the robustness of such detection models against missing sensor input (modality missing), e.g., due to a sudden sensor failure, is a critical problem which remains under-studied. In this work, we propose UniBEV, an end-to-end multi-modal 3D object detection framework designed for robustness against missing modalities: UniBEV can operate on LiDAR plus camera input, but also on LiDAR-only or camera-only input without retraining. To facilitate its detector head to handle different input combinations, UniBEV aims to create well-aligned Bird's Eye View (BEV) feature maps from each available modality. Unlike prior BEV-based multi-modal detection methods, all sensor modalities follow a uniform approach to resample features from the native sensor coordinate systems to the BEV features. We furthermore investigate the robustness of various fusion strategies w.r.t. missing modalities: the commonly used feature concatenation, but also channel-wise averaging, and a generalization to weighted averaging termed Channel Normalized Weights. To validate its effectiveness, we compare UniBEV to state-of-the-art BEVFusion and MetaBEV on nuScenes over all sensor input combinations. In this setting, UniBEV achieves $52.5 \%$ mAP on average over all input combinations, significantly improving over the baselines ($43.5 \%$ mAP on average for BEVFusion, $48.7 \%$ mAP on average for MetaBEV). An ablation study shows the robustness benefits of fusing by weighted averaging over regular concatenation, and of sharing queries between the BEV encoders of each modality. Our code will be released upon paper acceptance.

How Informative is the Approximation Error from Tensor Decomposition for Neural Network Compression?

Tensor decompositions have been successfully applied to compress neural networks. The compression algorithms using tensor decompositions commonly minimize the approximation error on the weights. Recent work assumes the approximation error on the weights is a proxy for the performance of the model to compress multiple layers and fine-tune the compressed model. Surprisingly, little research has systematically evaluated which approximation errors can be used to make choices regarding the layer, tensor decomposition method, and level of compression. To close this gap, we perform an experimental study to test if this assumption holds across different layers and types of decompositions, and what the effect of fine-tuning is. We include the approximation error on the features resulting from a compressed layer in our analysis to test if this provides a better proxy, as it explicitly takes the data into account. We find the approximation error on the weights has a positive correlation with the performance error, before as well as after fine-tuning. Basing the approximation error on the features does not improve the correlation significantly. While scaling the approximation error commonly is used to account for the different sizes of layers, the average correlation across layers is smaller than across all choices (i.e. layers, decompositions, and level of compression) before fine-tuning. When calculating the correlation across the different decompositions, the average rank correlation is larger than across all choices. This means multiple decompositions can be considered for compression and the approximation error can be used to choose between them.

Convolutional Cross-View Pose Estimation

We propose a novel end-to-end method for cross-view pose estimation. Given a ground-level query image and an aerial image that covers the query's local neighborhood, the 3 Degrees-of-Freedom camera pose of the query is estimated by matching its image descriptor to descriptors of local regions within the aerial image. The orientation-aware descriptors are obtained by using a translational equivariant convolutional ground image encoder and contrastive learning. The Localization Decoder produces a dense probability distribution in a coarse-to-fine manner with a novel Localization Matching Upsampling module. A smaller Orientation Decoder produces a vector field to condition the orientation estimate on the localization. Our method is validated on the VIGOR and KITTI datasets, where it surpasses the state-of-the-art baseline by 72% and 36% in median localization error for comparable orientation estimation accuracy. The predicted probability distribution can represent localization ambiguity, and enables rejecting possible erroneous predictions. Without re-training, the model can infer on ground images with different field of views and utilize orientation priors if available. On the Oxford RobotCar dataset, our method can reliably estimate the ego-vehicle's pose over time, achieving a median localization error under 1 meter and a median orientation error of around 1 degree at 14 FPS.

SliceMatch: Geometry-guided Aggregation for Cross-View Pose Estimation

This work addresses cross-view camera pose estimation, i.e., determining the 3-DoF camera pose of a given ground-level image w.r.t. an aerial image of the local area. We propose SliceMatch, which consists of ground and aerial feature extractors, feature aggregators, and a pose predictor. The feature extractors extract dense features from the ground and aerial images. Given a set of candidate camera poses, the feature aggregators construct a single ground descriptor and a set of rotational equivariant pose-dependent aerial descriptors. Notably, our novel aerial feature aggregator has a cross-view attention module for ground-view guided aerial feature selection, and utilizes the geometric projection of the ground camera's viewing frustum on the aerial image to pool features. The efficient construction of aerial descriptors is achieved by using precomputed masks and by re-assembling the aerial descriptors for rotated poses. SliceMatch is trained using contrastive learning and pose estimation is formulated as a similarity comparison between the ground descriptor and the aerial descriptors. SliceMatch outperforms the state-of-the-art by 19% and 62% in median localization error on the VIGOR and KITTI datasets, with 3x FPS of the fastest baseline.

How do Cross-View and Cross-Modal Alignment Affect Representations in Contrastive Learning?

Various state-of-the-art self-supervised visual representation learning approaches take advantage of data from multiple sensors by aligning the feature representations across views and/or modalities. In this work, we investigate how aligning representations affects the visual features obtained from cross-view and cross-modal contrastive learning on images and point clouds. On five real-world datasets and on five tasks, we train and evaluate 108 models based on four pretraining variations. We find that cross-modal representation alignment discards complementary visual information, such as color and texture, and instead emphasizes redundant depth cues. The depth cues obtained from pretraining improve downstream depth prediction performance. Also overall, cross-modal alignment leads to more robust encoders than pre-training by cross-view alignment, especially on depth prediction, instance segmentation, and object detection.