Fuzzy match index for scale-invariant feature transform (SIFT) features with application to face recognition with weak supervision

Seba Susan, Siddhant Jain, Abhishek Jain, Shikhar Verma, Aakash Sharma
  • IET Image Processing, November 2015, the Institution of Engineering and Technology (the IET)
  • DOI: 10.1049/iet-ipr.2014.0670

Fuzzy classifier for SIFT

What is it about?

A new fuzzy SIFT classifier is proposed that involves all the SIFT keypoints in the decision-making process. An entropy weight is introduced to highlight the contributions of some of the keypoints. To test the new fuzzy classifier, we developed an application for classifying uncropped faces against varying backgrounds, using a single training template. The entropy weight highlights the contributions of the facial features in this case.

Why is it important?

The novelty in our work is the weak supervision of the face recognition experiment using a single training template and using uncropped faces with different backgrounds. This is made possible through the fuzzy sift classifier proposed in our work that computes a fuzzy match index between the test and all the training templates.


Dr Seba Susan

I feel that this work marks an advancement over the present SIFT classification schemes as well as the weakly supervised face recognition task that requires minimum user intervention.

Read Publication


The following have contributed to this page: Dr Seba Susan

In partnership with: