Building a Multi-Source Imagery Dataset for 3D Reconstructions

This project entails building an extensive imagery dataset that can be used for developing 3D scene reconstruction methods in the field of Computer Vision. This novel and diverse dataset will consist of photographs, videos, drawings, paintings, and 3D scans of different sites on the UNCW campus. The overarching goal for this project is to design algorithms for aligning imagery of the same scene that humans can easily tell match but computers cannot.

Building a Multi-Source Imagery Dataset for 3D Reconstructions Presentation