2 September 2025 in IEEE Robotics and Automation Letters
Camera-based Bird's-Eye-View (BEV) representation has become a viable solution for 3D object detection in cost-effective autonomous driving. Currently, the explicit paradigm based on the lift-splat-shoot (LSS) pipeline has become one of the mainstream methods due to its efficiency and ease of deployment. However, the process of flattening the spatial representation in this pipeline mixes target features with excessive background noise. In addition, the generated BEV features are sparse due to the inherent characteristics of camera imaging. To address these limitations, we propose IAE-BEV, a novel two-stage multi-view 3D object detector that adaptively integrates instance features into BEV features, ultimately constructing a BEV representation that highlights instance information and alleviates sparsity. We also introduce the Occupancy Mask Pool, which enables instance features to interact with the 2D image plane in a more targeted and efficient manner. To further distinguish instances along the same camera ray, we design the Angle-Adaptive Self-Attention, which learns appropriate feature weights under the guidance of queries. Extensive experiments on the nuScenes dataset demonstrate the effectiveness and generalizability of our proposed framework, achieving up to +2.96% mAP improvement over state-of-the-art LSS-based baseline.
16 February 2026 in IEEE Transactions on Multimedia
Panoramic images have been used in various applications because of their ability to provide comprehensive spatial information. However, the high cost of obtaining panoramic images and the complexity of annotation pose serious obstacles to enhancing the performance of panoramic tasks by restricting the size and quality of datasets. The use of annotated planar images to synthesize panoramic images proves to be an effective approach to narrow this gap. Prior synthesis methods introduced new distortions that lead to inconsistent object shapes before and after synthesis. Concurrently, since the angle of view of the planar image is much smaller than that of the panoramic image, there is an issue of missing spatial information in the synthesized panoramic image. To address these challenges, we introduce a novel approach for converting planar images into panoramic ones with reduced deformation. For annotating targets in synthetic images, we develop a new algorithm based on determining the minimum spherical area to calculate spherical bounding boxes that closely adhere to object boundaries, rather than relying on an estimated target center point which results in inevitable calculation errors as in previous studies. Subsequently, we propose a new method for effectively filling any blank areas in the synthetic panoramic images to compensate for the loss of precision caused by absence of spatial information. Additionally, with a focus on the distortion characteristics of panoramic images, an innovative data augmentation strategy is devised to further enhance the model's ability to recognize objects in different positions. These methods are conducive to a more effective utilization of the rich planar image datasets for panoramic object detection tasks. In the experiments, we generated two synthetic panoramic datasets based on COCO for training models. Experimental results demonstrate that training with these synthetic datasets significantly improves prediction accuracy, far surpasses the state-of-the-art methods, and helps to unleash the full potential of panoramic object detection models. Source code is available at https://github.com/longlong-yu/official-panorama-coco.git.