Painting hand-drawn images using mobile devices

This thesis deals with the implementation of iOS application for colorization of hand-drawn images captured with the built-in camera. For image colorization the LazyBrush algorithm has been studied, implemented and tested. Requirements analysis, user interface design and usability testing was made in collaboration with intended target group of users—children. To obtain better visual results of colorization image preprocessing of captured image was implemented. Image captured with the built-in camera is not ideal for immediate colorization—it contains background, the picture may be rotated and perspectively disorted. These phenomena are removed through registration of paper edges, reverse perspective transformation using homography, image cropping and image filtration. This reduces the complexity of the workflow just to capture—colorize—save which makes the application usable easilly for children and unprofessional users.

Images

Attachments