基于改进YOLOv5的指纹二级特征检测方法

Fingerprint Second-Order Minutiae Detection Method Based on Improved YOLOv5

马德里指纹错案的出现使得现行指纹鉴定体系不断受到挑战和质疑。以指纹二级特征的统计规律为基础的量化鉴定技术成为了新的研究难点与热点,而指纹二级特征的自动检测与分类是实现指纹二级特征自动统计的基础。因此,提出一种基于YOLOv5的指纹二级特征检测方法。首先,建立指纹二级特征数据集,共包含4000张带标注的指纹图像;其次,根据指纹二级特征点尺寸小且分布密集的特点,对YOLOv5网络结构进行改进,删除原有的32倍下采样大目标特征检测层,添加新的微小特征融合层;再使用Feature Pyramid Networks(FPN)、Pyramid Attention Network(PAN)和Spatial Pyramid Pooling(SPP)结构通过融合多种尺度的方式实现局部特征和全局特征提取;最后,添加Squeeze-and-Excitation(SE)通道注意力机制模块,有效增强模型的鲁棒性和密集小目标的检测能力。实验结果表明,相比于原模型,改进后YOLOv5s_FI模型,在检测速度基本不变的情况下,平均精度均值(mAP0.5)从93.0%提高到97.4%,且权重缩减了3/4。

The fingerprint identification system has been challenged and questioned following the erroneous fingerprint individualization in the Madrid train bombings case.Therefore,the quantitative identification technology based on the statistical law of fingerprint secondary features is now a prevalent research topic,for which the automatic detection and classification of fingerprint secondorder minutiae serve as foundations.In this paper,a YOLOv5 based fingerprint secondorder minutiae detection method was proposed.First,a fingerprint secondorder minutiae dataset was established,which contained 4000 fingerprint images with annotations.The structure of the YOLOv5 network was improved based on the characteristics of small size and dense distribution of fingerprint secondorder minutiae.More specifically,the original feature detection layer of 32 times downsampled large target was deleted,and a new microscale detection layer was added.Feature Pyramid Networks(FPN),Pyramid Attention Network(PAN),and Spatial Pyramid Pooling(SPP)structures were used to extract local and global features through multiscale fusion.Finally,the SqueezeandExcitation(SE)channel attentional mechanism was added to effectively enhance the robustness of the model and the detection ability of dense small targets.The experimental results reveal that compared with the original model,the mean average precision(mAP0.5)value of the improved YOLOv5s_FI model increases from 93.0%to 97.4%under the condition that the detection speed is basically unchanged,and the weight of the improved YOLOv5s_FI model is reduced by three quarters.