UI Change Detect in Globalization: An Intelligent Screenshot-based Change Identification

Track: Technical | T9 |
Thursday, June 10, 2021, 10:00am – 10:45am
Held in: Jujama 2
Presenters:
Zhenyu Luo - VMware 
Daniel Ni - VMware

The solution starts with a pair of screenshots for a given webpage in any language. It divides the screenshots in a sequence of regions and performs a pixel level comparison. A red, green, blue-based subtraction indicates the difference between the compared regions and suggests the display change from one to another accordingly. By introducing optical character recognition extraction, it takes off the display text from the screenshots and fills the blank with the contextualizing color. The text-free comparison provides the focus on the identification of the layout change, like position and alignment of components with remarkable precision in practice.

Takeaways: Attendees will learn this approach brings users the vision on the display change of UI with incredibly high precision and speed. It changes the way of locale sensitive content verification. Users will be saving sufficient efforts with high satisfaction achieved by using this solution.