In recent times, there has been a notable surge in multimodal approaches that decorate raw LiDAR point clouds with camera-derived features to improve object detection performance. However, we found that these methods still grapple with the inherent sparsity of LiDAR point cloud data, primarily because fewer points are enriched with camera-derived features for sparsely distributed objects. We present an innovative approach that involves the generation of virtual LiDAR points using camera images and enhancing these virtual points with semantic labels obtained from image-based segmentation networks to tackle this issue and facilitate the detection of sparsely distributed objects, particularly those that are occluded or distant. Furthermore, we integrate a distance-aware data augmentation (DADA) technique to enhance the model’s capability to recognize these sparsely distributed objects by generating specialized training samples. Our approach offers a versatile solution that can be seamlessly integrated into various 3D frameworks and 2D semantic segmentation methods, resulting in significantly improved overall detection accuracy. Evaluation on the KITTI and nuScenes datasets demonstrates substantial enhancements in both 3D and bird’s eye view (BEV) detection benchmarks.

In this article, we present Sniffer Faster R-CNN++, an efficient camera-LiDAR late fusion network for low complexity and accurate object detection in autonomous driving scenarios. The proposed detection network architecture operates on output candidates of any three-dimensional (3D) detector and proposals from regional proposal network of any 2D detector to generate final prediction results. In comparison to the single modality object detection approaches, fusion-based methods in many instances suffer from dissimilar data integration difficulties. On the one hand, fusion-based network models are complicated in nature and, on the other hand, they require large computational overhead and resources, processing pipelines for training and inference specially, the early fusion and deep fusion approaches. As such, we devise a late fusion network that in-cooperates pre-trained, single-modality detectors without change, performing association only at the detection level. In addition to this, lidar-based method fail to detect distant object due to its sparse nature so we devise proposal refinement algorithm to jointly optimize detection candidates and assist detection for distant objects. Extensive experiments on both the 3D and 2D detection benchmark of challenging KITTI dataset illustrate that our proposed network architecture significantly improves the detection accuracy, accelerating the detection speed.

Autonomous vehicles need fast, reliable perception to detect vehicles, pedestrians, cyclists, traffic signs, and other road objects in real time. This paper surveys camera-based one-stage object detectors for autonomous driving. It traces 24 one-stage detectors, from YOLOv1 (2016) through SSD, RetinaNet, anchor-free designs such as FCOS and CenterNet, and efficiency-oriented models such as EfficientDet and RTMDet, to YOLOv12 (2025), with the real-time transformer RT-DETR as an end-to-end reference point. The detectors are compared by design choices, feature fusion, loss functions, and deployment trade-offs. Reported results are synthesized on general benchmarks (COCO, PASCAL VOC) and driving datasets (Waymo Open, KITTI, BDD100K, Argoverse-HD). The driving-dataset evidence, drawn mainly from single studies, suggests that accuracy and speed margins measured on general benchmarks may not carry over to road scenes. At high-input resolution on Waymo, ResNet-based one-stage detectors were not faster than Faster R-CNN, and small-object AP stayed at or below 10.3 for every one-stage model tested. Streaming evaluation on Argoverse-HD shows that accuracy drops once processing latency is taken into account. The survey also summarizes datasets, metrics, open challenges, and future directions, and argues that detectors should be evaluated on driving-specific data, under adverse conditions, and on target hardware before they are considered deployment-ready.

This research paper delves into the field of autonomous vehicle technology, examining the vulnerabilities inherent in each component of these transformative vehicles. Autonomous vehicles (AVs) are revolutionizing transportation by seamlessly integrating advanced functionalities such as sensing, perception, planning, decision-making, and control. However, their reliance on interconnected systems and external communication interfaces renders them susceptible to cybersecurity threats. This research endeavors to develop a comprehensive threat model for AV systems, employing OWASP Threat Dragon and the STRIDE framework. This model categorizes threats into Spoofing, Tampering, Repudiation, Information Disclosure, Denial of Service (DoS), and Elevation of Privilege. A systematic risk assessment is conducted to evaluate vulnerabilities across various AV components, including perception modules, planning systems, control units, and communication interfaces.