Webová aplikace pro vizuální úpravy fotografií s využitím generativní umělé inteligence

Abstract

This bachelor thesis deals with the design and implementation of a web application for visual photo editing using generative artificial intelligence. The user can upload an interior photo, crop it to the required size, manually draw a mask that marks the area to be edited and specify the desired modification using a text prompt. The application uses the OpenAI API for GPT Image 1.5 inpainting to regenerate the selected region of the image according to the prompt. Supabase is used for user management, storing before/after image pairs and handling application data, providing authentication, a PostgreSQL database and object storage. The thesis describes the technologies used, the design of the user interface, the communication with external services and evaluates the achieved results based on practical experiments.