Senior Design Projects

ECS193 A/B Winter & Spring 2022

Enhance a machine learning image analysis tool for skeletal muscle histological analysis

Name Lucas Smith
Affiliation UC Davis, Department of Neurobiology, Physiology, and Behavior

Project's details

Project title Enhance a machine learning image analysis tool for skeletal muscle histological analysis
Background Skeletal muscle makes up approximately 40% of body mass and is essential for human mobility, metabolic function, and overall health. Muscle fibers are the contractile cells that makeup muscle and some of the largest cells in the body. Muscle health is often evaluated based off the size and characteristics of the muscle fibers, which can undergo hypertrophy (fiber enlargement), atrophy (fiber shrinking), change in fiber type (fast or slow mechanical features), and central nucleation (indicative of regenerating fibers). Traditionally these parameters are evaluated by hand drawing fiber outlines and manual classification of fibers that is both tedious and unreliable. We have developed an image analysis software package designed specifically for skeletal muscle analysis that automates some of these procedures and is commonly used in skeletal muscle research, but has limitations that could be addressed.
Description The team working on this project will develop an updated version of the MATLAB-based image processing app for the analysis of muscle images. The app will be used by muscle researchers to automate the analysis of muscle cross sections using image segmentation techniques and machine learning for fast and reliable detection of muscle fibers, fiber types, regenerating fibers, and other non-fiber objects present in the images. The project team will design and build SMASH 2.0 including: • Improvement of the speed and accuracy of the current segmentation algorithm • Feature extraction and training of machine learning models • Building the app in an open source platform • Creating a GUI that is easy to use for a wide range of users • Testing and validating the accuracy of the app against manual methods
Deliverable • A working prototype of SMASH 2.0 • The code will be made available to the research community on a public depository • Students will assist with preparation of manuscripts and be co-authors on a scientific publication
Skill set desirable • Experience with MATLAB code • Java, C, C++, etc. • Image processing experience is a plus • An interest in biomedical research
Phone number 15307522930
Client time availability 30-60 min every two weeks
IP requirement Open source project
Attachment Click here
Selected Yes
Stuff N/A
TA Albara