Android app wireframes dataset

Description Dataset of 406 JSON files and JPG images, including the component name, size and position. These images were used for the development of the UIPAC model for automatic classification of a wireframe into a set of UI Design Patterns.
Data collection This dataset is composed of JSON files filtered from Sketch2aia (available under the creative commons Attribution-NonCommercial-ShareAlike 4.0 International license) and a part of RICO (copyright informations) using its UI Screenshots and hierachies with semantic annotations.
Data formats The dataset is grouped by class, each class is a folder that contains the instances.

These instances are in 2 formats, JSON and JPG, every instance has his representation in structural file (JSON) and image (JPG):
JSON: these files contains the size of the image file and describe all components contained in the screen with a name, position and size. There are 2 types of JSONs, the output of Sketch2aia and the output of RICO description, both containing the required component fields.

JPG: these files are images generated with the Python library PIL, creating a visual representation of the JSONs

Download Full dataset, organized into folders by each class
Git with the code of our model
Licence Our research is available under creative commons Attribution-NonCommercial-ShareAlike 4.0 International
Citation Schmitt Kreuch, E., Gresse von Wangenheim, C. And Carlo Rossa Hauck, J. (2021). User Interface structural components description dataset for the automatic classification of Wireframes into UI Design Patterns with Deep Learning, GQS/INCoD/INE/UFSC [Data set].
Keywords Wireframe, UI Design Patterns, App Inventor, Mobile application, Android, Deep learning