From 198309997cb6d98df1c17e0a222abe49664354f2 Mon Sep 17 00:00:00 2001
From: Guillaume Heusch <guillaumeheusch@Guillaumes-MacBook-Pro.local>
Date: Wed, 27 Jun 2018 14:54:53 +0200
Subject: [PATCH] [doc] updated the doc with the new scripts name
---
doc/guide_chrom.rst | 8 ++++----
doc/guide_cvpr14.rst | 16 ++++++++--------
doc/guide_performance.rst | 39 +++++++++++++++++++++++++++++++--------
doc/guide_ssr.rst | 8 ++++----
4 files changed, 47 insertions(+), 24 deletions(-)
diff --git a/doc/guide_chrom.rst b/doc/guide_chrom.rst
index e616258..8d27144 100644
--- a/doc/guide_chrom.rst
+++ b/doc/guide_chrom.rst
@@ -41,11 +41,11 @@ before the final pulse signal is built.
To extract the pulse signal from video sequences, do the following::
- $ ./bin/chrom_pulse.py config.py -vv
+ $ ./bin/bob_rppg_chrom_pulse.py config.py -vv
To see the full options, including parameters and protocols, type::
- $ ./bin/chrom_pulse.py --help
+ $ ./bin/bob_rppg_chrom_pulse.py --help
As you can see, the script takes a configuration file as argument. This
configuration file is required to at least specify the database, but can also
@@ -94,11 +94,11 @@ given below.
The execution of this script is very slow - mainly due to the face detection.
You can speed it up using the gridtk_ (especially, if you're at Idiap). For example::
- $ ./bin/jman sub -t 3490 -- ./bin/chrom_pulse.py cohface
+ $ ./bin/jman sub -t 3490 -- ./bin/bob_rppg_chrom_pulse.py cohface
The number of jobs (i.e. 3490) is given by typing::
- $ ./bin/chrom_pulse.py cohface --gridcount
+ $ ./bin/bob_rppg_chrom_pulse.py cohface --gridcount
.. _gridtk: https://pypi.python.org/pypi/gridtk
diff --git a/doc/guide_cvpr14.rst b/doc/guide_cvpr14.rst
index 9a23cd2..5dd152d 100644
--- a/doc/guide_cvpr14.rst
+++ b/doc/guide_cvpr14.rst
@@ -64,11 +64,11 @@ To extract the mean green colors the face region and of
the background across the video sequences of the defined database
in the configuration file, do the following::
- $ ./bin/cvpr14_extract_face_and_bg_signals.py config.py -vv
+ $ ./bin/bob_rppg_cvpr14_extract_face_and_bg_signals.py config.py -vv
To see the full options, including parameters and protocols, type::
- $ ./bin/cvpr14_extract_face_and_bg_signals.py --help
+ $ ./bin/bob_rppg_cvpr14_extract_face_and_bg_signals.py --help
Note that you can either pass parameters through command-line, or
by specififing them in the configuration file. Be aware that
@@ -80,11 +80,11 @@ the command-line overrides the configuration file though.
You can speed it up using the gridtk_ toolbox (especially, if you're at Idiap).
For example::
- $ ./bin/jman sub -t 3490 -- ./bin/cvpr14_extract_face_and_bg_signals. config.py
+ $ ./bin/jman sub -t 3490 -- ./bin/bob_rppg_cvpr14_extract_face_and_bg_signals. config.py
The number of jobs (i.e. 3490) is given by typing::
- $ ./bin/cvpr14_extract_signals.py cohface --gridcount
+ $ ./bin/bob_rppg_cvpr14_extract_face_and_bg_signals.py cohface --gridcount
Step 2: Illumination Rectification
@@ -96,7 +96,7 @@ Normalized Linear Mean Square and is then removed from the face signal. To get
the rectified green signal of the face area, you should execute the following
script::
- $ ./bin/cvpr14_illumination.py config.py -v
+ $ ./bin/bob_rppg_cvpr14_illumination.py config.py -v
Again, parameters can be passed either through the configuration file or
the command-line
@@ -113,8 +113,8 @@ channel on all the segment of all sequences. By default, the threshold is set su
of all the segments will be retained. To get the signals where large motion has
been eliminated, execute the following commands::
- $ ./bin/cvpr14_motion.py cohface --save-threshold threshold.txt -vv
- $ ./bin/cvpr14_motion.py cohface --load-threshold threshold.txt -vv
+ $ ./bin/bob_rppg_cvpr14_motion.py cohface --save-threshold threshold.txt -vv
+ $ ./bin/bob_rppg_cvpr14_motion.py cohface --load-threshold threshold.txt -vv
Step 4: Filtering
@@ -129,7 +129,7 @@ window. Finally, a bandpass filter is applied to restrict the
frequencies to the range corresponding to a plausible heart-rate. To filter the
signal, you should execute the following command::
- $ ./bin/cvpr14_filter.py cohface -vv
+ $ ./bin/bob_rppg_cvpr14_filter.py cohface -vv
A Full Configuration File Example
---------------------------------
diff --git a/doc/guide_performance.rst b/doc/guide_performance.rst
index b8984e6..8713371 100644
--- a/doc/guide_performance.rst
+++ b/doc/guide_performance.rst
@@ -13,12 +13,7 @@ signal. The Welch's algorithm is applied to find the power spectrum of the
signal, and the heart rate is found using peak detection in the frequency range
of interest. To obtain the heart-rate, you should do the following::
- $ ./bin/rppg_frequency_analysis.py hci -vv
-
-This script normally takes data from a directory called ``pulse``
-and outputs data to a directory called ``heart-rate``. This output represents
-the end of the processing chain and contains the estimated heart-rate for every
-video sequence in the dataset.
+ $ ./bin/bob_rppg_base_get_heart_rate.py config.py -v
Generating performance measures
@@ -27,7 +22,35 @@ Generating performance measures
In order to get some insights on how good the computed heart-rates match the
ground truth, you should execute the following script::
- $ ./bin/rppg_compute_performance.py hci --indir heart-rate -v -P
+ $ ./bin/bob_rppg_base_compute_performance.py config.py -v
This will output and save various statistics (Root Mean Square Error,
-Pearson correlation) as well as figures (error distribution, scatter plot)
+Pearson correlation) as well as figures (error distribution, scatter plot).
+
+
+Again, these scripts rely on the use of configuration
+files. An minimal example is given below:
+
+.. code-block:: python
+
+ import os, sys
+
+ import bob.db.hci_tagging
+ import bob.db.hci_tagging.driver
+
+ # DATABASE
+ if os.path.isdir(bob.db.hci_tagging.driver.DATABASE_LOCATION):
+ dbdir = bob.db.hci_tagging.driver.DATABASE_LOCATION
+ if dbdir == '':
+ print("You should provide a directory where the DB is located")
+ sys.exit()
+ database = bob.db.hci_tagging.Database()
+ protocol = 'cvpr14'
+
+ basedir = 'li-hci-cvpr14/'
+
+ # FREQUENCY ANALYSIS
+ hrdir = basedir + 'hr'
+ nsegments = 16
+ nfft = 8192
+
diff --git a/doc/guide_ssr.rst b/doc/guide_ssr.rst
index b993a26..d4237ab 100644
--- a/doc/guide_ssr.rst
+++ b/doc/guide_ssr.rst
@@ -36,11 +36,11 @@ After having applied the skin color filter, the full algorithm is applied,
as described in Algorithm 1 in the paper. To get the pulse signals for
all video in a database, do the following::
- $ ./bin/ssr_pulse.py config.py -v
+ $ ./bin/bob_rppg_ssr_pulse.py config.py -v
To see the full options, including parameters and protocols, type::
- $ ./bin/ssr_pulse.py --help
+ $ ./bin/bob_rppg_ssr_pulse.py --help
As you can see, the script takes a configuration file as argument. This
configuration file is required to at least specify the database, but can also
@@ -86,11 +86,11 @@ given below.
The execution of this script is very slow - mainly due to the face detection.
You can speed it up using the gridtk_ (especially, if you're at Idiap). For example::
- $ ./bin/jman sub -t 3490 -- ./bin/ssr_pulse.py cohface
+ $ ./bin/jman sub -t 3490 -- ./bin/bob_rppg_ssr_pulse.py cohface
The number of jobs (i.e. 3490) is given by typing::
- $ ./bin/ssr_pulse.py cohface --gridcount
+ $ ./bin/bob_rppg_ssr_pulse.py cohface --gridcount
.. _gridtk: https://pypi.python.org/pypi/gridtk
--
GitLab