Scale Invariant Feature Transform (SIFT) is one of the most widely used feature extraction algorithms to date. Its scale, translation, and rotation invariance, its robustness to change in contrast, brightness, and other transformations, make it the go-to algorithm for feature extraction and object detection. It was proposed by David Lowe in 2004.
The original publish paper can be found at http://www.cs.ubc.ca/~lowe/papers/ijcv04.pdf.
Some important properties of SIFT are as follows:
- It is invariant to scaling and rotation changes in objects
- It is also partially invariant to 3D viewpoints and illumination changes
- A large number of keypoints (features) can be extracted from a single image