The SIFT algorithm produces more keypoints but worse matches, leading to a denser point cloud.
The ORB algorithm, produces less keypoints but better matches, leading to a more sparse point cloud.
We chose to use the SIFT algorithm to have as many starting keypoints as possible and then applied our own custom filtering techniques.
Color based masking filters out parts of the image that are white while edge based masking finds the edge outline of the object and separates it from the background.
We chose to use edge based masking because it isolates the object cleanly without losing any parts of the object.