Skip to content

This repository contains the analysis I performed during the machine learning data challenge (MDI343) at Télécom Paris.

Notifications You must be signed in to change notification settings

camillecochener/Fusion-of-algorithms-for-face-recognition

Repository files navigation

Challenge Large Scale Machine Learning

Theme: fusion of algorithms for face recognition

Facial recognition is the process of identifying or verifying the identity of a person using their face. Today, it’s considered to be the most natural of all biometric measurements. Biometric means identity verification of persons according to their physical and behavioral characteristics. In the case of facial biometrics, a sensor captures a face, then transforms it into a vector of features and compares this transformed image to those held in a database. The biometric system confirms the correct match. The increasingly ubiquitous presence of biometric solutions and face recognition in particular in everyday life requires their adaptation for practical scenario. In the presence of several possible solutions, and if global decisions are to be made, each such single solution can be far less efficient than tailoring them to the complexity of an image.

Objective
The goal was to build a fusion of algorithms in order to construct the best suited solution for comparison of a pair of images. This fusion will be driven by qualities computed on each image.

The data
The data are from IDEMIA databases.

The training set includes 9.800.713 observations and 37 variables explicatives. The test set includes 3.768.311 predicted observations which is one-third of the training set.

The variables are as follows:

  • 13 variables relating to the quality of the second image (qr1 to qr13)
  • 11 matching scores between the two images (s1 to s11)
The match score is a measure of similarity between the first and second biometric feature vectors. When match scores output by different biometric matchers are consolidated in order to arrive at a final recognition decision, fusion is said to be done at the match score level.

About

This repository contains the analysis I performed during the machine learning data challenge (MDI343) at Télécom Paris.

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published