View Position
| Project Title | Optimal Path Planning and Steering Algorithms for Magnetic Resonance Imaging (MRI) Robotics |
| Summary | Magnetic resonance imaging (MRI) robotics is an emerging technology for minimally invasive intervention, automated robotic surgery, and therapeutic delivery. Micro/Nano robots in blood vessels can be driven and propelled by magnetic gradients with predefined path to deliver drugs or cells. To shorten steering distance, optimal path planning and steering algorithms are required to achieve targeted delivery quickly and accurately. In this project, the internship will focus on research of several path planning and steering algorithms in vascular network. Specifically, magnetic resonance angiography (MRA) image is preliminarily processed to generate blood vessels at first. Then, path planning method of A* algorithm, genetic algorithm, and neural network algorithm will be investigated and evaluated. The success of this project is able to identify short path planning and fast steering algorithm in real-time of MR scan. |
| Job Description | The intern is expected to: (1) Implement basic image processing algorithms on MRA images to extract vascular network. These basic algorithms include high-pass filtering, edge detection, and morphological dilation and erosion with Matlab toolbox. (2) Develop A* algorithm on vascular network. Given two points on blood vessels, the developed A* algorithm can identify the shortest path between two points or determine whether path connection between two points exists or not. (3) Investigate genetic algorithm based and neural network based path planning and steering algorithms. Similar to the above item (2), both genetic algorithm and neural network algorithm may be able to identify shortest path if it exists on vascular network. (4) Analyze results produced by three algorithms and evaluate their performance. Write a paper after complete the above work. |
| Use of Blue Waters | Blue waters super computing resources will be used in image processing of MRA images and path planning algorithms. Time complexity of A* algorithm, genetic algorithm, and neural network algorithm are super high, especially for MRA images. Without blue waters super computers, it is difficult to process and search targets in a short time. For example, A* algorithm searching on a MRA image with 100 x 100 pixels requires more than 16 hours on common computer. The proposed project will use larger MRA images with more pixels and dynamic 3D MRA sequence images, which require much more computation resources than processing a MRA image with 100 x 100 pixels. |
| Conditions/Qualifications | Must be familiar with Matlab programming. |
| Start Date | 05/31/2018 |
| End Date | 05/31/2019 |
| Location | University of Houston â Downtown, One Main St, Houston TX 77002 |
| Interns | He He
|