Eye

Independent Studies on Scale-Space Feature Extraction

I’ve taken a MSc in Signal Processing from National Technological University  (NTU) and one of the especially useful assignments there is the independent studies module where you can choose any topic to study. I’ve chosen Scale Space Feature extraction and it was quite eye opening on why  Laplacian of Gaussian (LoG) is able to produce features that are scale invariant. And then showing two implementation in the form of SIFT and SURF which both take different approximation techniques to reduce computation of the feature extraction.

ISM NTU
Tags: No tags

Leave a Comment