中国计量大学学报2018，Vol.29Issue(4)：398-405,8.DOI:10.3969/j.issn.2096-2835.2018.04.009
基于云平台和深度学习的软件GUI自动测试系统
Software GUI automatic testing systems based on the cloud platform and deep learning
摘要
Abstract
The graphical design of softwares graphical user interface (GUI) significantly influences the user experience.Testers have to do repetitive testing work without an established standard, which causes inaccurate testing scores of GUI and an inefficient work process.Aiming at the above problems, this paper constructs a GUI automatic testing system based on the cloud platform and uses the HOG+SVM model and AlexNet model respectively to extract the features of the GUI and classity it.In consideration of the small sample size of the GUI data, the strategy for improving the network performance of AlexNet using transfer learning was proposed.To fulfill users and algorithms requirements for computing performance, this system was deployed in a cloud platform.Users could real-timely evaluate softwares GUI on this system.Experiments showed that this system could run the GUI automatic testing with good performance, avoid the influence of the subjective factors of testers in the GUI testing process and reduce the workload of software testers.关键词
深度学习/迁移学习/GUI自动测试/云平台Key words
deep learning/transfer learning/GUI automatic testing/cloud platform分类
信息技术与安全科学引用本文复制引用
徐时怀,陆慧娟,叶敏超,严珂,金群..基于云平台和深度学习的软件GUI自动测试系统[J].中国计量大学学报,2018,29(4):398-405,8.基金项目
国家自然科学基金项目(No.61272315) （No.61272315）
浙江省科技计划项目(No.2017C34003) （No.2017C34003）
