基于轻量级神经网络的目标检测研究

Object detection research based on lightweight neural network

由于以CSPDarknet53为主干的YOLOv4神经网络参数量巨大,将其移植至手机等小型设备上时会降低其检测精度和速度,为了提高检测速度同时将检测精度控制在合理范围内,提出将原有的53层神经网络改为15层,并对其中的聚类算法进行优化,引入K-means++聚类算法对数据集进行分析,生成满足检测条件的Anchor Box;使用在负区间带有一定斜率的LeakyReLU激活函数代替存在梯度消失问题的Sigmoid激活函数,从而增强浅层网络的学习能力;同时考虑到Bounding Box与Anchor Box之间的中心距和宽高比具有一定的相关性,提出在原有损失函数的基础上增加相应的惩罚项生成L_(CIoU)损失函数,使损失函数在反向传播时梯度下降的方向性更好。实验结果表明,改进后的CSPDarknet15神经网络在VOC2007数据集上检测的平均精度达到83.94%,检测一幅图像的时间为3625 ms,与CSPDarknet53神经网络相比,检测速度提高了54.43%,能满足小型设备实时检测的速度和精度要求。

Due to the huge amount of parameters of the YOLOv4 neural network with CSPDarknet53 as the backbone,the detection accuracy and speed will be reduced when it is transplanted to small devices such as mobile phones.In order to improve the detection speed and control the detection accuracy within a reasonable range,this paper proposes to change the original 53-layer neural network to a 15-layer one,and optimizes its clustering algorithm.The K-means++clustering algorithm is introduced to analyze the data set to generate an anchor box that satisfies the detection conditions.LeakyReLU activation function with a certain slope in the negative interval is used to replace the Sigmoid activation function with vanishing gradients,thereby enhancing the learning ability of the shallow network.At the same time,considering that the center distance and the aspect ratio between the Bounding Box and the Anchor Box have a certain correlation,The corresponding penalty term is added to the original loss function to generate the L_(CIoU) loss function,so that the loss function has a better directionality of the gradient drop during back propagation.Experimental results show that the improved CSPDarknet15 neural network in the VOC2007 data set has an average detection accuracy of 83.94%,and the detection time of a picture is 3625 ms.Compared with the CSPDarknet53 neural network,the detection speed is increased by 54.43%,which can meet the speed and accuracy requirements of real-time detection of small devices.