Project Description:

Dynamic cardiovascular magnetic resonance imaging is crucial for diagnosing ischemic heart disease and detecting blood flow abnormalities that can lead to heart attacks. Deep learning (DL)-based reconstruction techniques like E2E-VarNet and CineVN are promising for speeding up scans, improving patient comfort, and enhancing diagnostic accuracy. However, DL models require large, diverse datasets to generalize well, and such data is scarce for cardiac MRI. This limits the development of robust DL models and increases their susceptibility to performance issues when applied to different scanners, contrast doses, or acquisition protocols. This project will explore the impact of dataset shifts (e.g., scanner type, contrast dose, acquisition protocols) on DL-based MRI reconstruction. We will use large datasets and advanced generative techniques, such as GANs and diffusion models, to improve model generalization across diverse conditions.

The student will design and implement neural network architectures, develop training strategies, and collaborate with medical imaging researchers to integrate clinical knowledge into the models. The student will also conduct experiments to benchmark the performance of new models against existing algorithms. Their work could lead to conference or journal publications, contributing to advancements in both medical imaging and deep learning.

Requirements:

Strong collaboration skills and the ability to effectively analyze research papers are essential. Proficiency in Python and experience with at least one major deep learning framework (such as PyTorch, TensorFlow, or JAX) are required. Knowledge of MATLAB is a plus. A good grasp of basic machine learning and deep learning concepts is also important. Research experience in deep learning, image processing, or computer vision would be a big plus. Knowledge of image processing, linear algebra, and statistics will also be a significant advantage.

Project Description:

The objective of this project is twofold. One of the objectives is to define a process using additive manufacturing of composite materials through which the body of an autonomous underwater vehicle (AUV) can be manufactured to meet mission specifics, actuation needs, and payload requirements. A modular vehicle will be designed for printing in sections with fiber-reinforced polymers using the Composites Additive Manufacturing Research Instrument (CAMRI) system developed at Purdue. Further, a joining technology will be developed to provide structural integrity and water resistance.

The second objective is to endow the system with motion capabilities through a control and navigation unit. This latter involves integrating a multiprocessor system with different sensors that can interface with the main computer through serial interfaces. The overall goal is to validate the navigation capabilities of the prototype and estimate its hydrodynamic properties through an experiment with a motion capture system at the dive well of Purdue.

This project is a collaboration between Dr. Nina Mahmoudian, Dr. Eduardo Barocio, and Dr. Jalil Chavez-Galaviz.

Requirements:

Required Skills:
-Controls
-Statics and Mechanics of Materials
-Machine and CAD Design
-Programming: Python, C, Matlab
-Communication Protocols: UART

Preferred Skills:
-FEA
-CNC Machining
-Commitment
-Teamwork
-Communication Protocols: Ethernet, SPI, IIC

Project Description:

The objective of this project is to investigate the process-structure-property relationship developed in non-planar printing with fiber reinforced thermoplastic polymers. The project involves using the multi-axis Composites Additive Manufacturing Research Instrument (CAMRI) developed at the Composites Manufacturing and Simulation Center at Purdue University. The student will learn to conduct experimental characterization of composite materials and to perform simulations of the printing process.

Requirements:

Required Skills/course work:
-Statics, Strength of Materials
-CAD Design
-Programming: Python, C, or Matlab
-Basic experience programming multi-axis industrial robots.

Project Description:

The goal of the project is to develop an intelligent platform that can be deployed in various bodies of water, such as local rivers and lakes, to measure critical environmental parameters. This platform can be used to support decision-making for aquaculture applications, such as algae or fish farming, or to provide data for assessing the Water Quality Index, aiding in the formulation of effective environmental policies.

For this purpose, we are leveraging a SOFAR Smart Mooring platform and Autonomous Surface Vehicles capable of collecting these parameters in real-time. This approach will enable continuous monitoring of water parameters while utilizing the collected data to dynamically plan sampling trajectories. By combining real-time data acquisition and adaptive path planning, our project aims to support the decision making in aquaculture and environmental policy making, providing insights for sustainable water resource management.

Requirements:

Controls, Electronics, Manufacturing (CNC machining, 3D printing), Programming