在目标检测领域中,基于交并比(intersection over union,IoU)的系列损失函数存在一定的局限性,使得边界框回归的精度和稳定性有待进一步提升.为此提出了一种基于非线性高斯平方距离的边界框回归损失函数.首先综合考虑了边界框中重叠性、中心点距离和长宽比3个因素,将边界框建模为高斯分布;然后提出一种高斯平方距离来衡量概率分布之间的差距;最后设计了符合优化趋势的非线性函数,将高斯平方距离转化为有利于神经网络学习的损失函数.实验结果表明,与IoU损失相比,所提方法在掩膜区域卷积神经网络、一阶全卷积目标检测器和自适应特征选择目标检测器上的平均精度均值分别提高了 0.3％、1.1％和2.3％,证明了该方法能有效提升目标检测的性能,同时有利于高精度边界框的回归.
Object Detection Based on Nonlinear Gaussian Squared Distance Loss
Existing series of loss functions based on intersection over union(IoU)have certain limitations,impacting the accuracy and stability of bounding box regression in object detection.To address this problem,a bounding box regression loss based on nonlin-ear Gaussian squared distance is proposed.Firstly,the three factors including overlapping,center point distance and aspect ratio in the bounding box are comprehensively considered,and the bounding box is modeled as a Gaussian distribution.Then a Gaussian squared distance is proposed to measure the distance between two distributions.Finally,a non-linear function is designed to transform the Gaussian square distance into a loss function that facilitates neural network learning.Experimental results show that compared with IoU loss,the mean average precision of the proposed method on mask region-based convo-lutional neural network,fully convolutional one-stage object detector and adaptive training sample selection object detector is improved by 0.3％,1.1％and 2.3％,respectively.These results demonstrate the efficiency of the proposed method in enhancing target detection performance and supporting the regression of high-precision bounding boxes.
object detectionbounding box regressionGaussian distributionintersection over union(IoU)convolutional neural network
万方数据
维普
