EagleVision: A Multi-Task Benchmark for Cross-Domain Perception in High-Speed Autonomous Racing

• URL: http://arxiv.org/abs/2604.11400v1
• Date: Mon, 13 Apr 2026 12:42:18 GMT
• Title: EagleVision: A Multi-Task Benchmark for Cross-Domain Perception in High-Speed Autonomous Racing
• Abstract summary: EagleVision is a unified LiDAR-based benchmark for 3D detection and trajectory prediction in high-speed racing.<n>It provides newly annotated 3D bounding boxes for the Indy Autonomous Challenge dataset and the A2RL Real competition dataset.<n>Using a dataset-centric transfer framework, we quantify cross-domain generalization across urban, simulator, and real racing domains.
• Score: 9.483261915968617
• License: http://creativecommons.org/licenses/by-nc-nd/4.0/
• Abstract: High-speed autonomous racing presents extreme perception challenges, including large relative velocities and substantial domain shifts from conventional urban-driving datasets. Existing benchmarks do not adequately capture these high-dynamic conditions. We introduce EagleVision, a unified LiDAR-based multi-task benchmark for 3D detection and trajectory prediction in high-speed racing, providing newly annotated 3D bounding boxes for the Indy Autonomous Challenge dataset (14,893 frames) and the A2RL Real competition dataset (1,163 frames), together with 12,000 simulator-generated annotated frames, all standardized under a common evaluation protocol. Using a dataset-centric transfer framework, we quantify cross-domain generalization across urban, simulator, and real racing domains. Urban pretraining improves detection over scratch training (NDS 0.72 vs. 0.69), while intermediate pretraining on real racing data achieves the best transfer to A2RL (NDS 0.726), outperforming simulator-only adaptation. For trajectory prediction, Indy-trained models surpass in-domain A2RL training on A2RL test sequences (FDE 0.947 vs. 1.250), highlighting the role of motion-distribution coverage in cross-domain forecasting. EagleVision enables systematic study of perception generalization under extreme high-speed dynamics. The dataset and benchmark are publicly available at https://avlab.io/EagleVision

Related papers

• Pedestrian Crossing Intent Classification From Event-Based Vision Using Convolutional Spiking Neural Networks With Temporal Augmentation [0.0]
  We present an end-to-end pipeline that converts real-world driving footage into synthetic vision sensor (DVS) event streams.<n>We also train a novel convolutional spiking neural network (Conv-SNN) with clip-consistent DVS augmentation to classify pedestrian crossing intent as binary.<n>The trained model achieves 95.83% accuracy and F1 = 0.9695 on the JAAD DVS test set, 97.79% accuracy and F1 = 0.9478 on normal CARLA DVS, and 94.78% accuracy and F1 = 0.8369 on adverse-weather CARLA DVS.
  arXiv  Detail & Related papers  (2026-09-11T02:34:24Z)
• Rethinking Pre-Training and Augmentation for Zero-Shot Cross-City Object Detection [14.031215272664538]
  We propose a new training pipeline for object detection structured around two core pillars.<n>Our framework bridges cross-city distribution gaps while using limited GPU memory (16GB)<n>Our optimized variants, RF-DETR-HR and RF-DETR-Grayworld, deliver a substantial empirical gain of +24.29 over the baseline.
  arXiv  Detail & Related papers  (2026-08-25T07:16:42Z)
• A2RL V\ extsubscript{max}: The A2RL autonomous racing dataset for long-range, high-speed perception and multi-vehicle interaction [39.79237723871247]
  This work introduces the A2RL Vtextsubscriptmax open-source perception dataset.<n>The dataset was captured during the 2024 Abu Dhabi Autonomous Racing League.<n>It contains almost 30,000 professionally annotated LiDAR point clouds, along with RADAR point clouds.
  arXiv  Detail & Related papers  (2026-07-20T10:59:32Z)
• FoundationGeo: Learning Spatial Pixel-Wise Fields for Monocular Metric Geometry [52.232826445242644]
  FoundationGeo bridges relative and metric prediction via spatial calibration and principled data design.<n>Stage 1 learns a high-fidelity, affine-invariant geometry model by initializing with DINOv3.<n>Stage 2 moves beyond global scaling by introducing pixel-wise calibration fields for metric estimation.
  arXiv  Detail & Related papers  (2026-07-13T14:10:01Z)
• CATS-V2V: A Real-World Vehicle-to-Vehicle Cooperative Perception Dataset with Complex Adverse Traffic Scenarios [25.181569981722976]
  We introduce CATS-V2V, the first-of-its-kind real-world dataset for V2V cooperative perception under complex adverse traffic scenarios.<n>The 100-clip dataset includes 60K frames of 10 Hz LiDAR point clouds and 1.26M multi-view 30 Hz camera images.<n>On this basis, we propose a target-based temporal alignment method, ensuring that all objects are precisely aligned across all sensor modalities.
  arXiv  Detail & Related papers  (2025-11-14T11:07:04Z)
• Detect Anything via Next Point Prediction [51.55967987350882]
  Rex- Omni is a 3B-scale MLLM that achieves state-of-the-art object perception performance.<n>On benchmarks like COCO and LVIS, Rex- Omni attains performance comparable to or exceeding regression-based models.
  arXiv  Detail & Related papers  (2025-10-14T17:59:54Z)
• Dual-Path Enhancements in Event-Based Eye Tracking: Augmented Robustness and Adaptive Temporal Modeling [0.0]
  Event-based eye tracking has become a pivotal technology for augmented reality and human-computer interaction.<n>Existing methods struggle with real-world challenges such as abrupt eye movements and environmental noise.<n>We introduce two key advancements. First, a robust data augmentation pipeline incorporating temporal shift, spatial flip, and event deletion improves model resilience.<n>Second, we propose KnightPupil, a hybrid architecture combining an EfficientNet-B3 backbone for spatial feature extraction, a bidirectional GRU for contextual temporal modeling, and a Linear Time-Varying State-Space Module.
  arXiv  Detail & Related papers  (2025-04-14T07:57:22Z)
• EMT: A Visual Multi-Task Benchmark Dataset for Autonomous Driving [8.97091577113286]
  Emirates Multi-Task dataset is designed to support multi-task benchmarking within a unified framework.<n>It comprises over 30,000 frames from a dash-camera perspective and 570,000 annotated bounding boxes, covering approximately 150 kilometers of driving routes.
  arXiv  Detail & Related papers  (2025-02-26T16:06:35Z)
• VECTOR: Velocity-Enhanced GRU Neural Network for Real-Time 3D UAV Trajectory Prediction [2.1825723033513165]
  We propose a new trajectory prediction method using Gated Recurrent Units (GRUs) within sequence-based neural networks. We employ both synthetic and real-world 3D UAV trajectory data, capturing a wide range of flight patterns, speeds, and agility. The GRU-based models significantly outperform state-of-the-art RNN approaches, with a mean square error (MSE) as low as 2 x 10-8.
  arXiv  Detail & Related papers  (2024-10-24T07:16:42Z)
• Transfer Learning from Simulated to Real Scenes for Monocular 3D Object Detection [9.708971995966476]
  This paper introduces a two-stage training strategy to address these challenges. Our approach initially trains a model on the large-scale synthetic dataset, RoadSense3D. We fine-tune the model on a combination of real-world datasets to enhance its adaptability to practical conditions.
  arXiv  Detail & Related papers  (2024-08-28T08:44:58Z)
• er.autopilot 1.0: The Full Autonomous Stack for Oval Racing at High Speeds [61.91756903900903]
  The Indy Autonomous Challenge (IAC) brought together nine autonomous racing teams competing at unprecedented speed and in head-to-head scenario, using independently developed software on open-wheel racecars. This paper presents the complete software architecture used by team TII EuroRacing (TII-ER), covering all the modules needed to avoid static obstacles, perform active overtakes and reach speeds above 75 m/s (270 km/h) Overall results and the performance of each module are described, as well as the lessons learned during the first two events of the competition on oval tracks, where the team placed respectively second and third.
  arXiv  Detail & Related papers  (2023-10-27T12:52:34Z)
• Semantic Segmentation in Satellite Hyperspectral Imagery by Deep Learning [54.094272065609815]
  We propose a lightweight 1D-CNN model, 1D-Justo-LiuNet, which outperforms state-of-the-art models in the hypespectral domain. 1D-Justo-LiuNet achieves the highest accuracy (0.93) with the smallest model size (4,563 parameters) among all tested models.
  arXiv  Detail & Related papers  (2023-10-24T21:57:59Z)
• Transforming Model Prediction for Tracking [109.08417327309937]
  Transformers capture global relations with little inductive bias, allowing it to learn the prediction of more powerful target models. We train the proposed tracker end-to-end and validate its performance by conducting comprehensive experiments on multiple tracking datasets. Our tracker sets a new state of the art on three benchmarks, achieving an AUC of 68.5% on the challenging LaSOT dataset.
  arXiv  Detail & Related papers  (2022-03-21T17:59:40Z)
• Real Time Monocular Vehicle Velocity Estimation using Synthetic Data [78.85123603488664]
  We look at the problem of estimating the velocity of road vehicles from a camera mounted on a moving car. We propose a two-step approach where first an off-the-shelf tracker is used to extract vehicle bounding boxes and then a small neural network is used to regress the vehicle velocity.
  arXiv  Detail & Related papers  (2021-09-16T13:10:27Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.