1

i succeded running openbr face recognition using command below.

$ br -algorithm FaceRecognition \
    -compare ../data/MEDS/img/S354-01-t10_01.jpg ../data/MEDS/img/S354-02-t10_01.jpg \
     -compare ../data/MEDS/img/S354-01-t10_01.jpg ../data/MEDS/img/S386-04-t10_01.jpg

i am getting output like this.

$ Set algorithm to FaceRecognition
$ Loading /usr/local/share/openbr/models/algorithms/FaceRecognition
$ Loading /usr/local/share/openbr/models/transforms//FaceRecognitionExtraction
$ Loading /usr/local/share/openbr/models/transforms//FaceRecognitionEmbedding
$ Loading /usr/local/share/openbr/models/transforms//FaceRecognitionQuantization
$ Comparing ../data/MEDS/img/S354-01-t10_01.jpg and ../data/MEDS/img/S354-02-t10_01.jpg
$ Enrolling ../data/MEDS/img/S354-01-t10_01.jpg to S354-01-t10_01r7Rv4W.mem
$ 100.00%  ELAPSED=00:00:00  REMAINING=00:00:00  COUNT=1
$ 100.00%  ELAPSED=00:00:00  REMAINING=00:00:00  COUNT=1
$ 1.8812
$ Comparing ../data/MEDS/img/S354-01-t10_01.jpg and ../data/MEDS/img/S386-04-t10_01.jpg
$ Enrolling ../data/MEDS/img/S354-01-t10_01.jpg to S354-01-t10_01r7Rv4W.mem
$ 100.00%  ELAPSED=00:00:00  REMAINING=00:00:00  COUNT=1
$ 100.00%  ELAPSED=00:00:00  REMAINING=00:00:00  COUNT=1
$ 0.571219

What is the meaning of values 1.8812,0.571219? From these values how can i know if faces are matching?

Miki
  • 40,887
  • 13
  • 123
  • 202
suraj kiran
  • 49
  • 1
  • 1
  • 9
  • I'm surprised about this too. Documentation lacks this explanation. Anyway, OpenBR works much worse then openface. Try openface: https://github.com/cmusatyalab/openface – Sharm Jan 31 '17 at 23:38
  • Were you able to figure this out? – Sadiq Ali Sep 08 '17 at 15:46

2 Answers2

0

Openbr evaluates by a scoring system how much an individual looks like another, and those values ​​should range from 0 to 1, closer to 1 means that the system achieved recognition, so that closer to 0 means that the response to recognition was bad. In this example, the program returned 1.8812 a value greater than 1, I think it should be a bug, but after testing quite openbr I realized that the higher the score, the greater the similarity between the comparison.

Danilo Cardoso
  • 1
  • But there is no information about the threshold that we could interpret as guarantee of 100% similarity of two persons. Sometimes two photos of the same person have rate 0.9 and sometimes 27. We able to use this values only for sorting to find the most similar persons on photos. –  Oct 11 '20 at 08:59
0

My experience:

Value Reliability
0 36.9%
0.1 50.0%
0.2 62.1%
0.3 71.8%
0.4 80.0%
0.5 85.2%
0.6 90.0%
1 97.4%
1.82 99.9%

Reliability is the probability that another person will not pass biometric identification.

vaskovsky
  • 1
  • Your answer could be improved with additional supporting information. Please [edit] to add further details, such as citations or documentation, so that others can confirm that your answer is correct. You can find more information on how to write good answers [in the help center](/help/how-to-answer). – Community Mar 21 '22 at 01:38