You are on page 1of 1

Human Face Analysis using Local Binary Pattern and

Persistence Homology
Face recognition is a well known field of study. The idea is to from an image of a
face, we extract the facial features and then use it tell whether two facial images are of the
same person or not. This essentially is the idea of face recognition. There are multiple
algorithms to perform this task. What we are trying to do here is that we are taking an existing
algorithm called Local Binary Pattern and introducing two new concepts to the algorithm,
namely Thresholding and Uniformily check. Our theory is that introduction of these two ideas
will further improve the algorithm. We analyse the results using another concept called
Persistence Homology. Persistence Homology gives us a complex structure called Simplicial
Complex. These complices can be used to recognize one’s face.

Simplicial complices are actually a mathematical concept. In mathematics, a


simplicial complex is a set composed of points, line segments, triangles, and their n-
dimensional counterparts. Persistence homology is a method used in topological data analysis
to study qualitative features of data that persist across multiple scales. There is one more
concept that is important to our cause which is Betti number. In algebraic topology, the Betti
numbers are used to distinguish topological spaces based on the connectivity of n-
dimensional simplicial complexes. A particular Betti number can identify some particular
things also like:
0th Betti number identifies no of components
1st Betti number identifies no of holes
2nd Betti number identifies no of voids/cavities, etc

Some other concepts that are at use are Bar code diagram, persistence diagram etc.

To fulfill this project, we had to design algorithms regarding noise reduction, edge
creation, creation of simplicial complices etc. We have analyzed the final outcome of the
project and looked into what further modifications can be done to improve. Therefore, there is
still many ways to improve the project, but still we have a significant outcome from our work.

You might also like