The goal of this project is for you to gain experience with methods in computer vision that are too complicated to be applied in an assignment. You are free to pick any topic which interests you, has a significant computer vision component and represents effort consistent with the percentage of your grade assigned to this project. You must apply your work to actual sensor image data (not simulated). You can use code off the web, but you must attribute it, and provide significant extension or analysis to make your work appropriate for this project.
The project description should be a brief (approximately 2 page) document describing the problem you plan to address and the approach you plan to take. You should briefly introduce the problem, explain why it is interesting or important, describe the current state of the art, and tell me what your project will contribute. I expect you to cite a few (≥5) relevant papers from the literature. Email me a pdf copy of the description. I will contact you if I feel your proposal needs modification.
The final report will be a complete description of your project. It will include a thorough explanation of your method, implementation and findings. Please submit an electronic version (pdf) along with any relevant code, images and data. You will be expected to give a 15 minute presentation of your work in class in early Dec.
Contact me early and often if you have problems!!!
Pointers you may find useful: