Vehicle main and sub-brand identification method based on multi-layer network of attention mechanism

The invention belongs to the field of computer vision and the technical field of intelligent traffic information, and particularly relates to a vehicle main and sub-brand identification method based on a multi-layer network of an attention mechanism, which can obtain a main and sub-brand identification result of a vehicle at the same time. According to an attention mechanism-based multi-layer network vehicle main and sub-brand identification model, more attention is paid to a salient region through an attention mechanism, and the generalization ability of the model is improved; and through a voting mechanism, the model is more stable and reliable in a complex scene. 本发明属于计算机视觉领域和智能交通信息技术领域，具体涉及一种基于注意力机制的多层网络的车辆主子品牌识别方法，本发明的基于注意力机制的多层网络的车辆主子品牌识别方法，能同时得到车辆的主子品牌识别结果；基于注意力机制的多层网络的车辆主子品牌识别模型通过注意力机制，更加关注显著性区域，提高了模型的泛化能力；通过投票机制，使得模型在复杂场景下更加稳定和可靠。