基于卷积神经网络的深度学习技术在软件缺陷检测中的应用  

Application of Deep Learningin Software Defect Detection Based on Convolutional Neural Networks

作　　者：胡韬 杨阳 Hu Tao;Yang Yang(Yunnan Technology and Business University,Kunming 651701,China)

机构地区：[1]云南工商学院,昆明651701

出　　处：《黑龙江科学》2024年第14期146-148,共3页Heilongjiang Science

摘　　要：探讨了卷积神经网络(CNN)在软件缺陷检测中的应用。采用深度学习技术,模拟图像识别中的模式识别能力,对代码进行自动分析,以识别潜在缺陷。实验结果显示,该方法的缺陷检测正确率达到了94.28%~97.51%,说明利用CNN进行软件缺陷检测能够有效提升检测速度和准确性,对于降低开发成本、提高软件质量及可靠性具有重要意义。The study discusses the application of convolutional neural network(CNN)in software defect detection,simulates model recognition capabilities in image recognition with deep learning techniques,and automatically analyzes the codes to identify potential defects.Experiments show that the correct rate of defect detection system is 94.28%~97.51%.The use of CNN to detect software defects can effectively improve the detection speed and accuracy,and is of great significance for reducing development costs and improving software quality and reliability.

关 键 词：深度学习 应用软件 缺陷检测 卷积神经网络 系统框架 

分 类 号：TP391.1[自动化与计算机技术—计算机应用技术] TP183[自动化与计算机技术—计算机科学与技术]