Extreme value machine: An algorithm for open set classification

Executive Summary

Open set classification is a problem of identifying/classifying unknown classes. Unknown classes are images used neither in training nor in testing. With recent advent of Deep learning based techniques, all the conventional classification techniques are replaced with deep learning algorithms. However, Deep Learning (DL) based algorithm does not perform well in open set classification.

In this section, we will discuss about open set classification and difficulties in using DL approaches. We will also see the formulation of Extreme value machine algorithm, state of art algorithm used for open set classification problem. Python implementation of Extreme value machine (EVM) algorithm is available open source , but it is difficult to use python implementation in android applications. eInfochips recreated this algorithm on Java platform to use it in a face recognition android application.

Project Highlights

Extreme value machine: An algorithm for open set classification
  • Introduction
  • Known space vs Unknown space
  • Facial recognition pipeline
  • Simple classifier
  • Statistical Extreme values
  • Extreme value theorem
  • Extreme Value Machine formulation
  • Probability matrix computation
  • Extreme vectors and model selection
  • New feature probability calculation
  • Applications
  • Conclusion
To Download This Resource
Fill in the details below
By submitting this form, I acknowledge that I have read and understand the Privacy Policy
I wish to be contacted by eInfochips

Our Work





Device Partnerships
Digital Partnerships
Quality Partnerships
Silicon Partnerships


Products & IPs