Dr. Mahdi Champour (Jampour)

Developing advanced AI techniques to reveal undertexts in palimpsests, producing reconstructions that are contextually clearer and more readable than original MSI images. This groundbreaking work significantly enhances document preservation and content recovery.
The main objectives of this project are to explore the feasibility of revealing undertexts in palimpsests and to integrate image inpainting techniques for reconstructing undertexts in MSI images using Generative AI. By applying undertext image inpainting, the project aims to produce reconstructions that are contextually clearer, more readable, and more useful for experts compared to the original MSI images. more
Related publications: Revealing Palimpsests with Latent Diffusion Models, and Character Localization in Degraded Historical Documents

My team and I researched and developed a comprehensive, end-to-end intelligent system focused on pistachio processing. By employing computer vision techniques and advanced deep learning models, the system accurately detects and classifies pistachios based on specific characteristics, including type, size, color, and quality, contributing significantly to the optimization and precision of agricultural processing.

My team and I designed an interactive platform to detect and analyze children’s body movements and reactions, incorporating them as characters within the game. Projected onto the floor or wall, this setup allows children to engage physically and actively within the game environment. The project emphasized promoting physical activity, interactive learning, and cognitive exercise for children, proving especially beneficial during the restrictive COVID-19 period.

In this research we proposed a cutting-edge deep learning algorithms tailored for colorizing historical grayscale images and videos. By proposing advanced restoration techniques, we achieved vivid colorization, breathing life back into historical visual data and enabling a refreshed interpretation of archival materials.
Related publications: Advanced multi-GANs, and Multi-GANs

I invented and developed an intelligent librarian robot, namely Pars, for shelf-reading using computer vision approach to maintain the library resources. The Pars robot demonstrates advanced capabilities, including detecting, recognizing, and interpreting books to assess whether specific resources require attention. With intelligent navigation, it identifies and reports misplaced books, offering substantial support for organized library management.
Related publications: Shelf-reader Robot, and Decoding Blurred Barcodes (Persian)
In media: Iran Science Watch, and IRIB News Agency