基于语义识别和机器视觉的智能审核方法研究

Research on Intelligent Audit Method based on Semantic Recognition and Machine Vision

为实现报告自动化审核,本文提出通过基于网页提取信息填充报告模板,根据脚本算法审核报告基本信息,将报告中的文字及图片内容进行特征提取以及深度学习,识别报告结论的逻辑及技术准确性。结合优化OpenCV图像识别算法,在实现图片占用最小空间特征识别的基础上,实现图片信息提取。该研究具有较好的通用性,为实验室报告审核工作提供高效解决思路。

In order to realize the automatic review of reports,this paper proposes to fill in the report template by extracting information based on the web page,review the basic information of the report according to the script algorithm,extract the features of the text and pictures in the report and conduct in-depth learning to identify the logical and technical accuracy of the report conclusion.Combined with the optimized OpenCV image recognition algorithm,the image information extraction is realized on the basis of the recognition of the minimum space occupied by the image.The research has good generality,and provides an efficient solution for the report review of all laboratories in CATARC.