Commit c03edc27 authored by André Anjos's avatar André Anjos 💬

Merge branch 'issue_3' into 'master'

Issue 3



See merge request !15
parents 27e49e53 445c2a0c
Pipeline #5004 passed with stages
in 19 minutes and 59 seconds
; vim: set fileencoding=utf-8 :
; Manuel Guenther <manuel.guenther@idiap.ch>
; Thu Oct 9 16:51:06 CEST 2014
[buildout]
......
......@@ -42,10 +42,6 @@ Usually it is a good idea to have at least verbose level 2 (i.e., calling
specifies the number of parallel jobs you want to execute.
In the remainder of this section we introduce baseline experiments you can
readily run with this tool without further configuration. Baselines examplified
in this guide were published in [TVM14]_.
Database setups and baselines are encoded using
:ref:`bob.bio.base.configuration-files`, all stored inside the package root, in
the directory ``bob/bio/vein/configurations``. Documentation for each resource
......@@ -86,6 +82,11 @@ is available on the section :ref:`bob.bio.vein.resources`.
Once this step is done, you can proceed with the instructions below.
In the remainder of this section we introduce baseline experiments you can
readily run with this tool without further configuration. Baselines examplified
in this guide were published in [TVM14]_.
Repeated Line-Tracking with Miura Matching
==========================================
......@@ -106,6 +107,15 @@ protocol, do the following:
submit your job for SGE execution, you can run it in parallel (using 4
parallel tasks) by adding the options ``--parallel=4 --nice=10``.
Optionally, you may use the ``parallel` resource configuration which already
sets the number of parallel jobs to the number of hardware cores you have
installed on your machine (as with :py:func:`multiprocessing.cpu_count`) and
sets ``nice=10``. For example:
.. code-block:: sh
$ ./bin/verify.py verafinger rlt parallel -vv
This command line selects and runs the following implementations for the
toolchain:
......@@ -124,7 +134,7 @@ performance:
.. code-block:: sh
$ ./bin/bob_eval_threshold.py --scores <path-to>/verafinger/rlt/Nom/nonorm/scores-dev --criterium=eer
$ ./bin/bob_eval_threshold.py <path-to>/verafinger/rlt/Nom/nonorm/scores-dev
('Threshold:', 0.32045327)
FAR : 26.362% (12701/48180)
FRR : 26.364% (58/220)
......@@ -157,7 +167,7 @@ we obtained:
.. code-block:: sh
$ ./bin/bob_eval_threshold.py --scores <path-to>/verafinger/mc/Nom/nonorm/scores-dev --criterium=eer
$ ./bin/bob_eval_threshold.py <path-to>/verafinger/mc/Nom/nonorm/scores-dev
('Threshold:', 0.078274325)
FAR : 3.182% (1533/48180)
FRR : 3.182% (7/220)
......@@ -191,7 +201,7 @@ we obtained:
.. code-block:: sh
$ ./bin/bob_eval_threshold.py --scores <path-to>/verafinger/wld/NOM/nonorm/scores-dev --criterium=eer
$ ./bin/bob_eval_threshold.py <path-to>/verafinger/wld/NOM/nonorm/scores-dev
('Threshold:', 0.239141175)
FAR : 10.455% (5037/48180)
FRR : 10.455% (23/220)
......@@ -202,23 +212,37 @@ Results for other Baselines
===========================
This package may generate results for other combinations of protocols and
databases. Here is a summary table for some variants (results are expressed the
the equal-error rate on the development set, in percentage):
======================== ================= ====== ====== ====== ======
Approach UTFVP Vera
------------------------------------------ ------------- -------------
Feature Extractor Post Processing B Full B Full
======================== ================= ====== ====== ====== ======
Maximum Curvature Histogram Eq.
Maximum Curvature None 3.2
Repeated Line Tracking Histogram Eq.
Repeated Line Tracking None 26.4
Wide Line Detector Histogram Eq. 8.2
Wide Line Detector None 10.4
======================== ================= ====== ====== ====== ======
WLD + HEQ (preproc) @ Vera/Full = 10.9%
databases. Here is a summary table for some variants (results expressed
correspond to the the equal-error rate on the development set, in percentage):
======================== ================= ====== ====== ====== ====== ======
Approach Vera Finger UTFVP
------------------------------------------ -------------------- -------------
Feature Extractor Post Processing Full B Nom 1vsall nom
======================== ================= ====== ====== ====== ====== ======
Repeated Line Tracking None 23.9 24.1 24.9 1.7 1.4
Repeated Line Tracking Histogram Eq. 26.2 23.6 24.9 2.1 0.9
Maximum Curvature None 3.2 3.2 3.1 0.4 0.
Maximum Curvature Histogram Eq. 3.0 2.7 2.7 0.4 0.
Wide Line Detector None 10.2 10.2 10.5 2.3 1.7
Wide Line Detector Histogram Eq. 8.0 9.7 7.3 1.7 0.9
======================== ================= ====== ====== ====== ====== ======
In a machine with 48 cores, running these baselines took the following time
(hh:mm):
======================== ================= ====== ====== ====== ====== ======
Approach Vera Finger UTFVP
------------------------------------------ -------------------- -------------
Feature Extractor Post Processing Full B Nom 1vsall nom
======================== ================= ====== ====== ====== ====== ======
Repeated Line Tracking None 01:16 00:23 00:23 12:44 00:35
Repeated Line Tracking Histogram Eq. 00:50 00:23 00:23 13:00 00:35
Maximum Curvature None 03:28 00:54 00:59 58:34 01:48
Maximum Curvature Histogram Eq. 02:45 00:54 00:59 49:03 01:49
Wide Line Detector None 00:07 00:01 00:01 02:25 00:05
Wide Line Detector Histogram Eq. 00:04 00:01 00:01 02:04 00:06
======================== ================= ====== ====== ====== ====== ======
.. include:: links.rst
py:class bob.bio.base.database.BioFile
py:class bob.bio.base.database.BioDatabase
py:class BioFile
py:class bob.bio.base.database.database.BioDatabase
py:class BioFileSet
py:class bob.bio.base.database.file.BioFile
py:class bob.bio.base.preprocessor.Preprocessor.Preprocessor
py:class bob.bio.base.extractor.Extractor.Extractor
py:class bob.bio.base.algorithm.Algorithm.Algorithm
py:class BioFile
py:class BioFileSet
# These are not documented at all in Python 2.7, but works for 3.x
py:exc ValueError
Markdown is supported
0%
or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment