Commit 232f4a21 authored by André Anjos's avatar André Anjos

Merge branch 'issue-17' into 'master'

Using the new bob.bio API

Closes #17

See merge request !43
parents 0693a692 93c487b6
Pipeline #20783 passed with stages
in 34 minutes and 32 seconds
......@@ -147,11 +147,18 @@ performance:
.. code-block:: sh
$ bob_eval_threshold.py <path-to>/verafinger/rlt/Nom/nonorm/scores-dev
('Threshold:', 0.31835292)
FAR : 23.636% (11388/48180)
FRR : 23.636% (52/220)
HTER: 23.636%
$ bob bio metrics <path-to>/verafinger/rlt/Nom/nonorm/scores-dev --no-evaluation
[Min. criterion: EER ] Threshold on Development set `scores-dev`: 0.31835292
====== ========================
None Development scores-dev
====== ========================
FtA 0.0%
FMR 23.6% (11388/48180)
FNMR 23.6% (52/220)
FAR 23.6%
FRR 23.6%
HTER 23.6%
====== ========================
Maximum Curvature with Miura Matching
......@@ -180,11 +187,19 @@ we obtained:
.. code-block:: sh
$ bob_eval_threshold.py <path-to>/verafinger/mc/Nom/nonorm/scores-dev
('Threshold:', 0.0737283)
FAR : 4.388% (2114/48180)
FRR : 4.545% (10/220)
HTER: 4.467%
$ bob bio metrics <path-to>/verafinger/mc/Nom/nonorm/scores-dev --no-evaluation
[Min. criterion: EER ] Threshold on Development set `scores-dev`: 7.372830e-02
====== ========================
None Development scores-dev
====== ========================
FtA 0.0%
FMR 4.4% (2116/48180)
FNMR 4.5% (10/220)
FAR 4.4%
FRR 4.5%
HTER 4.5%
====== ========================
Wide Line Detector with Miura Matching
======================================
......@@ -192,7 +207,7 @@ Wide Line Detector with Miura Matching
You can find the description of this method on the paper from Huang *et al.*
[HDLTL10]_.
To run the baseline on the `VERA fingervein`_ database, using the ``NOM``
To run the baseline on the `VERA fingervein`_ database, using the ``Nom``
protocol like above, do the following:
......@@ -213,11 +228,17 @@ we obtained:
.. code-block:: sh
$ bob_eval_threshold.py <path-to>/verafinger/wld/NOM/nonorm/scores-dev
('Threshold:', 0.240269475)
FAR : 9.770% (4707/48180)
FRR : 9.545% (21/220)
HTER: 9.658%
$ bob bio metrics <path-to>/verafinger/wld/Nom/nonorm/scores-dev --no-evaluation
[Min. criterion: EER ] Threshold on Development set `scores-dev`: 2.402707e-01
====== ========================
None Development scores-dev
====== ========================
FtA 0.0%
FMR 9.8% (4726/48180)
FNMR 10.0% (22/220)
FAR 9.8%
FRR 10.0%
HTER 9.9%
Results for other Baselines
......@@ -233,7 +254,7 @@ correspond to the the equal-error rate on the development set, in percentage):
Feature Extractor Full B Nom 1vsall nom
======================== ====== ====== ====== ====== ======
Repeated Line Tracking 14.6 13.4 23.6 3.4 1.4
Wide Line Detector 5.8 5.6 9.7 2.8 1.9
Wide Line Detector 5.8 5.6 9.9 2.8 1.9
Maximum Curvature 2.5 1.4 4.5 0.9 0.4
======================== ====== ====== ====== ====== ======
......@@ -347,11 +368,11 @@ When used to run an experiment,
:py:class:`bob.bio.vein.preprocessor.WatershedMask` requires you provide a
*pre-trained* neural network model that presets the markers before
watershedding takes place. In order to create one, you can run the program
`markdet.py`:
`bob_bio_vein_markdet.py`:
.. code-block:: sh
$ markdet.py --hidden=20 --samples=500 fv3d central dev
$ bob_bio_vein_markdet.py --hidden=20 --samples=500 fv3d central dev
You input, as arguments to this application, the database, protocol and subset
name you wish to use for training the network. The data is loaded observing a
......@@ -367,9 +388,9 @@ Region of Interest Goodness of Fit
==================================
Automatic region of interest (RoI) finding and cropping can be evaluated using
a couple of scripts available in this package. The program ``compare_rois.py``
compares two sets of ``preprocessed`` images and masks, generated by
*different* preprocessors (see
a couple of scripts available in this package. The program
``bob_bio_vein_compare_rois.py`` compares two sets of ``preprocessed`` images
and masks, generated by *different* preprocessors (see
:py:class:`bob.bio.base.preprocessor.Preprocessor`) and calculates a few
metrics to help you determine how both techniques compare. Normally, the
program is used to compare the result of automatic RoI to manually annoted
......@@ -379,7 +400,7 @@ extracted ones. E.g.:
.. code-block:: sh
$ compare_rois.py ~/verafinger/mc_annot/preprocessed ~/verafinger/mc/preprocessed
$ bob_bio_vein_compare_rois.py ~/verafinger/mc_annot/preprocessed ~/verafinger/mc/preprocessed
Jaccard index: 9.60e-01 +- 5.98e-02
Intersection ratio (m1): 9.79e-01 +- 5.81e-02
Intersection ratio of complement (m2): 1.96e-02 +- 1.53e-02
......@@ -399,12 +420,12 @@ metrics.
Pipeline Display
================
You can use the program ``view_sample.py`` to display the images after
full processing using:
You can use the program ``bob_bio_vein_view_sample.py`` to display the images
after full processing using:
.. code-block:: sh
$ ./bin/view_sample.py --save=output-dir verafinger /path/to/processed/directory 030-M/030_L_1
$ bob_bio_vein_view_sample.py --save=output-dir verafinger /path/to/processed/directory 030-M/030_L_1
$ # open output-dir
And you should be able to view images like these (example taken from the Vera
......@@ -415,7 +436,7 @@ feature extractor):
:scale: 50%
Example RoI overlayed on finger vein image of the Vera fingervein database,
as produced by the script ``view_sample.py``.
as produced by the script ``bob_bio_vein_view_sample.py``.
.. figure:: img/binarized.*
......
......@@ -50,11 +50,11 @@ setup(
],
'console_scripts': [
'compare_rois.py = bob.bio.vein.script.compare_rois:main',
'view_sample.py = bob.bio.vein.script.view_sample:main',
'blame.py = bob.bio.vein.script.blame:main',
'markdet.py = bob.bio.vein.script.markdet:main',
'watershed_mask.py = bob.bio.vein.script.watershed_mask:main',
'bob_bio_vein_compare_rois.py = bob.bio.vein.script.compare_rois:main',
'bob_bio_vein_view_sample.py = bob.bio.vein.script.view_sample:main',
'bob_bio_vein_blame.py = bob.bio.vein.script.blame:main',
'bob_bio_vein_markdet.py = bob.bio.vein.script.markdet:main',
'bob_bio_vein_watershed_mask.py = bob.bio.vein.script.watershed_mask:main',
]
},
......
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