摘要
随着人工智能(AI)技术的快速发展,神经网络算法在各领域不断突破,我们期待神经网络算法可以更好的应用到笔迹领域并绽放光彩。由书写习惯可变性与稳定性可知,同源笔迹其少量特征发生变化的同时其主要特征是稳定的。本文通过高效液相色谱法(条件:柱温:35℃,检测波长:580 nm)与笔迹特征几何量化分别获取笔墨理化数据与几何形态数据,利用BP神经网络误差逆向传播算法分析笔迹特征,通过神经网络进行反复训练、学习得到合适的权值、阈值,为后续笔迹检验提供数据支持。在传统笔迹检验领域与理化分析中引入计算机辅助,使笔迹检验更客观、科学。
With the rapid development of artificial intelligence(AI)technology,neural network algorithm in various fields continue to break through,we expect neural network algorithm can be better applied to the field of handwriting and bloom brilliantly.According to the variability and stability of writing habits,the main features of homologous handwriting are stable while a small number of features change.In this paper,through high performance liquid chromatography(condition:column temperature:35℃,detection wavelength:580 nm)and geometric quantization of handwriting features,the physical and chemical data and geometric shape data of pen and ink are obtained respectively.The BP neural network error back propagation algorithm is used to analyze the handwriting features.The neural network is used to repeatedly train and learn to obtain appropriate weights and thresholds,which provides data support for subsequent handwriting inspection.In the field of traditional handwriting inspection and physical and chemical analysis,the introduction of computer assistance makes handwriting inspection more objective and scientific.
作者
赵彦涵
Zhao Yanhan(Criminal Investigation Police University of China,Shenyang 110035,China)
出处
《山东化工》
CAS
2021年第2期150-152,共3页
Shandong Chemical Industry
