On October 26, 2019, the second NCTS Summit, hosted by Testin, with the theme of "AI+Future", was officially held at the Beijing International Convention Center. At the summit, Testin President Xu Kun officially launched iTestin, a new AI product of Testin Cloud Testing, a testing business. Xu Kun personally demonstrated the amazing performance of AI+testing on iTestin using natural language, which amazed everyone. Testers using iTestin only need to enter natural language text descriptions such as click, wait, check, long press, and input in the interface, and the AI in the background can accurately and real-time perform complete and correct operations and related instructions on the real machine. AI technology support As an application of Internet infrastructure, cloud testing is an important part of the application life cycle, and cloud testing is also integrating and promoting AI technology innovation to further improve enterprise production efficiency and release enterprise productivity. Unlike traditional test scripts that require a lot of manual intervention, script recording and maintenance are achieved by manually selecting target controls; iTestin's new natural language script greatly reduces the difficulty of script recording and greatly improves product usability and script maintenance efficiency. These are due to Cloud Testing's years of technical development and deep learning technology. Lowering the threshold and increasing efficiency Unlike traditional script test writing, which requires a lot of code support and professional knowledge, testers using iTestin only need to enter natural language text descriptions such as click, wait, check, long press, and input in the interface, and the background AI can accurately and in real time perform complete and correct operations and related instructions on the real machine. iTestin is also the first testing tool based on natural language; it has a script recording function, and also supports natural language interaction and automated batch deployment operations; this greatly reduces the threshold for testers to use and improves the efficiency of test script recording. At the specific operation level, for example, if the user only needs to enter "click to search", iTestin can automatically detect and identify the search icon on the current page of the App and complete the click operation of the icon. Traditional test scripts require a lot of code support and testing, while iTestin's new natural language script greatly reduces the difficulty of script recording. While greatly improving the ease of use of test products and the efficiency of script maintenance, it also greatly reduces the threshold and technical difficulty of testers. Efficient image and text recognition engine based on deep learning The difficulty iTestin encountered during the research process was the recognition of mobile phone apps. In order to solve the icon recognition problem, Testin created a high-efficiency engine for OCR text recognition and icon recognition based on deep learning. As we all know, the foundation of machine learning is the accumulation of a large amount of data. According to reports, based on the scenario characteristics of the App, Testin specially collected more than 200,000 App UI screenshots, plus the tens of millions of data accumulated by Testin's customers over the years, annotated with millions of words and icons, and combined with the data accumulated by Cloud Testing, completed the training of the tens of millions of data set model in the testing industry. Based on a powerful deep learning model, the text recognition accuracy of Testin OCR model exceeds 97%, which comprehensively surpasses the high-precision OCR services provided by other Internet companies, and has reached the industry's high level in key performance indicators such as text recognition accuracy and algorithm operation speed. In addition, Testin's icon recognition engine has achieved a recognition accuracy of more than 95% for common icons such as "search, share, return, menu", etc. At the same time, Xu Kun said that Yunce will continue to increase its investment and efforts in this regard, and looks forward to being able to recognize icons that even humans cannot recognize in the future. Editor's message: The launch of iTestin is a great event for the testing market. While bringing good news to a large number of testers, it is also gradually promoting AI+testing to become a consensus in the industry. I believe that the implementation of artificial intelligence in the testing industry will usher in a new era for the entire testing industry. 【51CTO.com original article】 [51CTO original article, please indicate the original author and source as 51CTO.com when reprinting on partner sites] 【Editor: Captain TEL: (010) 68476606】 |
<<: Application practice of cross-platform mobile framework UniApp
The basic meaning of seo is search engine optimiz...
Product retention is similar to product onboardin...
Nowadays, many companies and enterprises are cond...
Compared with the "heart", "tone&q...
Everyone knows that people with high blood pressu...
This article is going to talk about: Why are more...
[[149874]] Clint Hocking, the designer of Splinte...
APP is a word that has only appeared in our lives...
Apple added the App Tracking Transparency feature...
Kunming tea tasting has its own studio. Recommend...
Since the development of the mobile game market, ...
【Kangge English】Practical English resources of Ch...
Any society coexists with its information and med...
According to reports, US President Trump stated a...
In the early stages of a business , when you are ...