Front Vehicle Detection Method Based on Multi-scale Fusion and Attention
XU Ding
Abstract：This paper proposes a neural network method based on multi-scale feature fusion and channel attention mechanism to accurately detect the front vehicle.The MSCA-Y model proposed in this article combines two key insights based on YOLOv4.This paper proposes an efficient multi-scale feature fusion network.On the basis of making full use of the feature representation ca-pability of the backbone network,the spatial information of multi-scale feature maps is effectively fused to enhance the performance of detecting micro-vehicles.This paper introduces the channel attention mechanism,by strengthening the attention to the character-istics of key parts of the vehicle's various postures,to further obtain better detection performance in complex detection tasks.In or-der to prove the effectiveness of the method,this paper evaluates the model on the KITTI dataset.Experiments show that this method achieves a higher mAP on the benchmark of the KITTI dataset(reaching 87.12%in the difficult subset).
Keywords：front vehicle detectionmulti-scale fusionchannel attentionconvolutional neural network
Publication Date：2023-11-20
Online Publishing Date：2025-08-15(First online date of this platform, not the publication date of the document)
Pages：7( 2573-2579 )