3d Classification and Reconstruction
This project is a pipe line project that starts from identifying and masking interesting objects to creating a 3d mesh of these masked interesting object, after which they are aligned to the original image and finally the entire object masked in the scene are all aligned back to create a 3d reconstruction of the 2d scene. The project has a head count of about 367 people and some of the quality measure include, drawing an accurate mask around an object, not masking more than one object, selecting the best mesh during mesh section, selecting the mesh with the best colour distribution and aligning the mesh to the original image in both 2d and 3d views while maintaining the gravity alignment, and colinear of objects.