Two images of the same bear on a desk taken at different horitzonal positions relative to the subject were taken and anaylzeed using the SIFT algorithm implemented in Matlab with the aid of a toolbox by VLFeat.

Figure 4.1. Picture A

Taken on a DSLR Nikon D3000

Figure 4.2. Picture B

A horizontal shift from camera position in Picture A

After the SIFT algorithm has been applied corresponding features are very clear and intuitive.

Figure 4.3. Zoomed In Feature Extraction Picutre A

Features(Green) with scale and orientation

Figure 4.4. Zoomed In Feature Extraction Picutre B

Features(Green) with scale and orientation

It is clear that the SIFT program chooses the same features of the same scale and orientation in each image, and there is a clear match between the two images.

Figure 4.5. Matching Features

Black line drawn from center of corresponding features.