This work was done as part of the final project for a computer vision course offered at the Johns Hopkins University. We used two images of Maryland Hall, a building on the Homewood campus at the Johns Hopkins University, to reconstruct a 3D point cloud of the building. We used SIFT to find and match feature points in the two images while rejecting outliers, and triangulated the matched points to obtain the 3D reconstruction (SNH11). Our results were voted the best CV project video for 3D reconstruction by everyone present at the demos.
[SNH11] A Sinha, Hau Sze, Diane Duros. “3D Reconstruction of a Hopkins Landmark”, Submitted as final project report for EN.600.461(01): Computer Vision, The Johns Hopkins University, Baltimore, MD (December 10, 2011)
Best Computer Vision Project Video for 3D Reconstruction, 2011