Title: Automated segmentation of metal corrosion using deep learning
Student: Stuart, Jake email@example.com
Advisor: Jelena Tešić
Date and Time: Thu Dec 8 2022 1:00 PM
Abstract: The inspection of for corrosion in metal is laborious, slow, and often requires a human in the loop. Recently, deep learning-based algorithms have revealed promise and performance in the automatic detection of corrosion. However, to date, research regarding the segmentation of images for automated corrosion detection has been limited, due to the lack of availability of training data and corrosion annotation masks. There have been several attempts in industrial engineering to leverage existing metal corrosion data to perform automated corrosion image segmentation in infrastructure inspection e.g. multi-step approach to determine if corrosion is present before attempting to create a pixel-level corrosion segmentation mask. In this project, we apply the similar paradigm to study metal corrosion due to biofilm formation under microgravity and full gravity conditions in imagery provided by McLean Lab @ TXST. While we wait for the annotation data, we demonstrate our deep learning segmentation pipeline on few open source datasets.
Deadline: Dec. 23, 2022, midnight