Papers

NameDesriptionRelease DateDownload LinksEvent
How much real data do we actually need: Analyzing object detection performance using synthetic and real dataSupervised training of deep models requires a large amount of annotated data to be available. However, data annotation is a tremendously exhausting and costly task to perform. One alternative is to use synthetic data. This paper provides a comprehensive study of the effects of replacing real data with synthetic data.Jul 2019https://arxiv.org/abs/1907.07061International Conference on Machine Learning, ICML 2019 - AI for Autonomous Driving, Long Beach, California, USA
Deep Open Space Segmentation using Automotive RadarThis paper describes an AI approach for advanced deep segmentation from radar. This neural network is trained to identify open space in parking scenarios. Different deep models are evaluated with various radar input representations. This system achieves low memory usage and real-time processing speeds, and is thus very well suited for embedded deployment.Mar 2020https://arxiv.org/pdf/2004.03449.pdfInternational Conference on Microwaves for Intelligent Mobility, ICMIM 2020, Linz, Austria
PolarNet: Accelerated Deep Open Space Segmentation Using Automotive Radar in Polar DomainCamera and Lidar processing have been revolutionized with the rapid development of deep learning model architectures. Automotive radar is one of the crucial elements of automated driver assistance and autonomous driving systems. Radar still relies on traditional signal processing techniques, unlike camera and Lidar based methods. We believe this is the missing link to achieve the most robust perception system. Identifying drivable space and occupied space is the first step in any autonomous decision making task. Occupancy grid map representation of the environment is often used for this purpose. In this paper, we propose PolarNet, a deep neural model to process radar information in polar domain for open space segmentation. We explore various input-output representations. Our experiments show that PolarNet is a effective way to process radar data that achieves state-of-the-art performance and processing speeds while maintaining a compact size.Mar 2021https://arxiv.org/abs/2103.033877th International Conference on Vehicle Technology and Intelligent Transport Systems (VEHITS 2021)
Point Cloud based Hierarchical Deep Odometry EstimationProcessing point clouds using deep neural networks is still a challenging task. Most existing models focus on object detection and registration with deep neural networks using point clouds. In this paper, we propose a deep model that learns to estimate odometry in driving scenarios using point cloud data. The proposed model consumes raw point clouds in order to extract frame-to-frame odometry estimation through a hierarchical model architecture. Also, a local bundle adjustment variation of this model using LSTM layers is implemented. These two approaches are comprehensively evaluated and are compared against the state-of-the-art.Mar 2021https://arxiv.org/abs/2103.033947th International Conference on Vehicle Technology and Intelligent Transport Systems (VEHITS 2021)

Datasets

NameDesriptionRelease DateDownload LinksAssociated Paper
SCORPA publically available dataset of radar observations called SCORP was collected. It is composed of 3913 frames, collected in 11 driving sequences. It is annotated with all drivable open spaces in the scene and accompanied with corresponding camera images The dataset gives access to Analog-to-Digital Converter (ADC) radar signals and annotations.Mar 2020 Dowload https://arxiv.org/pdf/2004.03449.pdf