Graphical User Interface (GUI) Dataset

In the era of smart devices, users increasingly demand smooth and intelligent interactive interfaces. However, existing user interface designs mostly rely on traditional static analysis, which struggles to meet the requirements of dynamic user interaction and complex, changing scenarios. Many smart devices cannot accurately predict user intentions in practical applications, leading to issues such as choppy interactions. This dataset aims to enhance the intelligent interaction capabilities of user interfaces and improve user experience through video collection and analysis. Data collection uses high-definition camera equipment to record user operations in a variety of real-world application scenarios. Quality control is ensured through multiple rounds of annotation, consistency checks, and reviews by professional human-computer interaction experts, ensuring high-quality data. The annotation team consists of 30 experts with experience in human-computer interaction design. Data preprocessing includes video frame extraction, deduplication, and normalization processing to ensure data consistency and applicability. The final data is stored in MP4 format and organized by operational steps and scenario classification for efficient and easy usage. The data is stored on a cloud platform in a distributed manner for convenient access and training.