Final project report: FINAL-REPORT.md

This repository contains code and documentation for Rust 101 project - Sebastian Petrik (Plasmoxy#7521), Igor Durica (duricaigor#0164), Andrii Rybak (Kiligram#6228).

Project proposal

We would like to implement an GUI/server application using various technologies, focusing mainly on image manipulation including cropping, inverting, detecting faces, taking images from camera and other interesting features that'll come to our minds.

We hope to explore/learn:

• Rust's modular system - separate different functionalities into standalone modules giving the ability to reuse them as libraries and allowing splitting the work between members
• Integration of different libraries/technologies together
  • integrating our functionalties into a web server application / into a GUI application
  • integration with a deep learning library for face detection in the photos
• Custom image manipulation implementation (for example, inverting pixel colors to produce a negative) - possibly with Rayon, or using an existing image manipulation crate
• Concurrency and parallelism in Rust for our features (using multiple cpu threads for image processing, asynchronous server endpoints implementation)

We hope to implement these features:

• Build a GUI inferface for our functionalities - either with native Rust fraemwork or Webview-based framework - https://tauri.app/
• Build a REST API for our functionalities using a web server framework
• Selected image manipulation functionalities (for example, crop an image, invert vertically/horizontally, set aspect ratio, invert colors)
• Identifying face on photo and cropping the person (using deep learning model OR just using a Rust crate implementation)
• Taking photo with camera (for now without live preview, simply take photo)

Possible dependencies:

• Tauri - web frontend / rust backend - framework for GUI
• Rocket or Axum https://github.com/tokio-rs/axum - backend framework
• Tokio - for concurrency with rocket.rs
• Rayon - for (possible) parallelisation of our image procesing functionality
• Tensorflow - loading and inference of a face recognition DL model, or possibly even Tract (learn to leverage vast ecosystem of universal ONNX DL model inference)
• Nokhwa - taking a photo with a camera