EMIFF：面向车路协同三维目标检测的增强型多尺度图像特征融合
EMIFF: Enhanced Multi-scale Image Feature Fusion for Vehicle-Infrastructure Cooperative 3D Object Detection
AI总结：
本文提出EMIFF框架，通过多尺度交叉注意力、相机感知通道掩码和特征压缩模块，解决车路协同三维检测中的相机不同步位姿误差与带宽受限问题，在DAIR-V2X-C上达到SOTA。
摘要：
在自动驾驶中，协同感知利用来自车辆和基础设施的多视角相机，提供超越单一车辆视角的全局视野和丰富的道路语义上下文。目前，车路协同三维（VIC3D）目标检测面临两大挑战：1）由于各相机之间的时间不同步，融合多视角图像时存在固有位姿误差；2）有限的通信带宽导致传输过程中的信息损失。为解决这些问题，我们提出了一种新颖的基于相机的VIC3D任务三维检测框架——增强型多尺度图像特征融合（EMIFF）。为充分利用车辆和基础设施的整体视角，我们提出了多尺度交叉注意力（MCA）和相机感知通道掩码（CCM）模块，在尺度、空间和通道层面增强基础设施与车辆特征，以纠正由相机不同步引入的位姿误差。我们还引入了一个包含通道压缩块和空间压缩块的特征压缩（FC）模块，以提高传输效率。实验表明，EMIFF在DAIR-V2X-C数据集上达到了SOTA性能，在传输成本相当的情况下显著优于此前的早期融合和晚期融合方法。
英文摘要：
In autonomous driving, cooperative perception makes use of multi-view cameras from both vehicles and infrastructure, providing a global vantage point with rich semantic context of road conditions beyond a single vehicle viewpoint. Currently, two major challenges persist in vehicle-infrastructure cooperative 3D (VIC3D) object detection: $1)$ inherent pose errors when fusing multi-view images, caused by time asynchrony across cameras; $2)$ information loss in transmission process resulted from limited communication bandwidth. To address these issues, we propose a novel camera-based 3D detection framework for VIC3D task, Enhanced Multi-scale Image Feature Fusion (EMIFF). To fully exploit holistic perspectives from both vehicles and infrastructure, we propose Multi-scale Cross Attention (MCA) and Camera-aware Channel Masking (CCM) modules to enhance infrastructure and vehicle features at scale, spatial, and channel levels to correct the pose error introduced by camera asynchrony. We also introduce a Feature Compression (FC) module with channel and spatial compression blocks for transmission efficiency. Experiments show that EMIFF achieves SOTA on DAIR-V2X-C datasets, significantly outperforming previous early-fusion and late-fusion methods with comparable transmission costs.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
