用多表示融合提升雷达目标检测效率与精度
Multi-Representation Adapter with Neural Architecture Search for Efficient Range-Doppler Radar Object Detection
- 将雷达图同时表示为热力图和灰度图，捕获高层语义与细节纹理
- 设计适配器分支与交换模块，实现多表示特征的高效提取与融合
- 通过神经架构搜索优化模型结构，兼顾高精度与低计算开销
相比摄像头，雷达传感器在恶劣光照和天气条件下更具鲁棒性，因而基于雷达的目标检测日益受到关注。本文提出一种针对距离-多普勒（RD）雷达图的高效目标检测模型。首先，将RD雷达图以热力图和灰度图像两种形式表示，以获取高层语义特征和细粒度纹理信息。随后，设计额外的适配器分支、支持双模式的交换模块以及主-辅融合模块，分别实现特征的有效提取、跨表示特征交换与融合。此外，构建包含多种宽度与融合操作的超网络，并采用单次神经架构搜索方法，进一步提升模型效率同时保持高性能。实验结果表明，该模型在准确率与效率之间取得良好平衡。在RADDet和CARRADA数据集上分别达到71.9和57.1的mAP@50，刷新当前最佳性能。
原文摘要 · Abstract (English)
Detecting objects efficiently from radar sensors has recently become a popular trend due to their robustness against adverse lighting and weather conditions compared with cameras. This paper presents an efficient object detection model for Range-Doppler (RD) radar maps. Specifically, we first represent RD radar maps with multi-representation, i.e., heatmaps and grayscale images, to gather high-level object and fine-grained texture features. Then, we design an additional Adapter branch, an Exchanger Module with two modes, and a Primary-Auxiliary Fusion Module to effectively extract, exchange, and fuse features from the multi-representation inputs, respectively. Furthermore, we construct a supernet with various width and fusion operations in the Adapter branch for the proposed model and employ a One-Shot Neural Architecture Search method to further improve the model's efficiency while maintaining high performance. Experimental results demonstrate that our model obtains favorable accuracy and efficiency trade-off. Moreover, we achieve new state-of-the-art performance on RADDet and CARRADA datasets with mAP@50 of 71.9 and 57.1, respectively.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。