Handheld Device Doodle Art Dataset

The "Handheld Device Doodle Art Dataset" comprises of 171,137 hand-drawn doodles, that have 169 classes, it was initially developed as a part of a drawing app that we were working on. These images were created on mobile screens on an app developed on Flutter using it's canvas library, made specifically to create data. The process engaged a cooperative group of 20 friends, it is also taken care that each data element are drawn with at least 5 different styles and data is balanced. As a resource, this dataset offers significant potential for training of machine learning models in hand drawing recognition specifically those investigating human drawing behaviour on handheld devices. Furthermore, this dataset could prove instrumental in the development of drawing practice applications for children.