Incorrect Chi-Square distance used in LBPH FaceRecognizer (Bug #3514)


Added by Ivy Lee about 11 years ago. Updated about 11 years ago.


Status:Done Start date:2014-01-30
Priority:Normal Due date:
Assignee:Ivy Lee % Done:

0%

Category:legacy, contrib
Target version:3.0
Affected version:branch 'master' (3.0-dev) Operating System:Linux
Difficulty: HW Platform:x64
Pull request:https://github.com/Itseez/opencv/pull/2237

Description

When using LBPH face recognizer, in prediction, the algorithm computes the Chi-Square distance between two spatial histograms, the correct Chi-Square distance should be CV_COMP_CHISQR_ALT introduced in this commit:
https://github.com/vkocheganov/opencv/commit/086db9d6db73956be2ee18b8a552a9179da6b6fd

Reference paper regarding LBPH: Ahonen, T., Hadid, A., and Pietikainen, M. Face Recognition with Local Binary Patterns. Computer Vision - ECCV 2004 (2004), 469–481.


Associated revisions

Revision b7f2b671
Added by Ivy Lee about 11 years ago

Fixed Bug #3514 Incorrect Chi-Square distance used in LBPH FaceRecognizer

Revision 03df7831
Added by Vadim Pisarevsky about 10 years ago

Merge pull request #3514 from mshabunin:compat-24

History

Updated by Ivy Lee about 11 years ago

  • Assignee set to Ivy Lee

Updated by Vladislav Vinogradov about 11 years ago

Hello Ivy,

Thank you for reporting the problem.

If you could work on a solution for the issue on your side, a contribution (see http://www.code.opencv.org/projects/opencv/wiki/How_to_contribute) would be very appreciated!

  • Category set to legacy, contrib
  • Status changed from New to Open

Updated by Vladislav Vinogradov about 11 years ago

  • Pull request set to https://github.com/Itseez/opencv/pull/2237

Updated by Vladislav Vinogradov about 11 years ago

  • Affected version changed from 2.4.8 (latest release) to branch 'master' (3.0-dev)

Updated by Vladislav Vinogradov about 11 years ago

The pull request was merged into master branch.
Thanks you for contribution!

  • Target version set to 3.0
  • Status changed from Open to Done

Also available in: Atom PDF