Building a fullstack mobile application with flutter and stable diffusion model

Abstract

The thesis’ core focus is on the development and implementation of a mobile application using Flutter and the Stable Diffusion model. TikTok has grown as a dominating social network in today’s digital world, particularly among the younger generation, overtaking even giants like Facebook and Instagram. Creators often create short, entertaining video content immediately from their mobile devices, which contributes to the platform’s appeal. Furthermore, artificial intelligence has grown in popularity, particularly since the release of ChatGPT near the end of 2022. The thesis finds its niche, trying to investigate the concept of directly generating images using artificial intelligence on mobile devices.

The thesis is divided into three parts. The first section presents a thorough technology overview and explains fundamental artificial intelligence models, establishing the groundwork for readers. The second section focuses on the planning and development of the mobile application, which includes fundamental functionality for viewing and generating Japanese drawing-style graphics from real-life photos. It provides useful information about the technical aspects of front-end and back-end implementations. The thesis finishes with a thorough overview that includes recommendations for improving mobile applications.

In short, this thesis is a helpful resource for people interested not just in Flutter and the Dart programming language, but also in artificial intelligence models in the context of image generation, specifically the Stable Diffusion model. By combining these features, the thesis demonstrates how to develop and implement a full-stack solution for a mobile application that is meant to function flawlessly on Android smartphones.