Development of client server AI services with Mac OS

• Main Author: Ng, Alexander Rui Hao
• Other Authors: Chen Change Loy
• Format: Final Year Project
• Language: English
• Published: Nanyang Technological University 2024
• Subjects: Computer and Information Science; CoreML; Full stack web application; Vapor; Python flask; Image editor; React
• Online Access: https://hdl.handle.net/10356/175313
• Institution: Nanyang Technological University

This project is a full stack application similar to Runway ML, a popular platform that offers Artificial Intelligence services to assist with creative projects such as image generation and image editing. The application developed in this project provides the following services - Background Removal, Super Resolution, Image Generation and Image Search and the result images can be added seamlessly to a canvas where it can be moved, resized and combined with other images to form a project. The front-end is a React application which provides a canvas where images can be added. Artificial Intelligence assisted image editing services can then be applied to the images on the canvas. The back-end consists of 2 parts. Firstly, a Core ML server built on Vapor where Core ML models are deployed as RestAPIs for Background Removal, Super Resolution and Image Generation. Secondly, a CLIP server built on Python Flask for generating text and image feature embeddings which are stored on a Solr database. Image searching is performed using a KNN vector search on the feature embeddings.