In order to run for image blobs at multiple scale, the SIFT may construct a scale space, defined as follows. Code the Gaussian kernel. Thus in practice the scale space is arranged by. Blobs are detected as local extrema of the Difference of Gaussians Mmorristown resolution space, obtained by subtracting successive scales of the Basic scale space:..
Notice that the scale level index s varies in a slightly redundant set. Once extracted, local extrema are quadratically interpolated this is very important especially at the lower resolution scales in order to have accurate keypoint localization at the full resolution. Finally, they are filtered to eliminate low-contrast responses or responses close to edges and the orientation s are assigned, as explained next. Peaks which are too short may have been generated by noise and are discarded.