bob issueshttps://gitlab.idiap.ch/groups/bob/-/issues2018-04-26T17:02:50Zhttps://gitlab.idiap.ch/bob/bob.ip.gabor/-/issues/4the Transform class does not allow to specify arbitrary wavelets2018-04-26T17:02:50ZManuel Günthersiebenkopf@googlemail.comthe Transform class does not allow to specify arbitrary waveletsAs discussed in https://groups.google.com/forum/#!topic/bob-devel/y6PVFkD_V68 the `Transform` class cannot use an arbitrary selection of wavelets, but always fills with a grid of wavelets. It would be great to have the possibility to pas...As discussed in https://groups.google.com/forum/#!topic/bob-devel/y6PVFkD_V68 the `Transform` class cannot use an arbitrary selection of wavelets, but always fills with a grid of wavelets. It would be great to have the possibility to pass an `[Wavelet]` to the constructor, too.
This would include to go over all the code and see, where the wavelets are actually used. I guess, it will also need a change in the I/O, i.e., to write and read the set of wavelets when they are set by hand.Manuel Günthersiebenkopf@googlemail.comManuel Günthersiebenkopf@googlemail.comhttps://gitlab.idiap.ch/bob/bob.db.mnist/-/issues/5Images are returned in wrong format (black and white exchanged)2018-04-19T05:51:10ZManuel Günthersiebenkopf@googlemail.comImages are returned in wrong format (black and white exchanged)I have just come across a weired thing with respect to the image representation of the MNIST images. When reading the instructions on the MNIST web page: http://yann.lecun.com/exdb/mnist/ it is written there that:
> Pixel values are 0 t...I have just come across a weired thing with respect to the image representation of the MNIST images. When reading the instructions on the MNIST web page: http://yann.lecun.com/exdb/mnist/ it is written there that:
> Pixel values are 0 to 255. 0 means background (white), 255 means foreground (black).
which is exactly the opposite to how images are usually stored: black is 0 and white is 255.
So, what do you think, should we change the `images` to be `255-images` in https://gitlab.idiap.ch/bob/bob.db.mnist/blob/929cbe3bec9c8a04cc08202e1b7ce15e1fe86871/bob/db/mnist/query.py#L60, or shall we keep it as it is (since we are not really converting them to images here)? Shall we at least mention this in our documentation?https://gitlab.idiap.ch/bob/bob.db.verafinger/-/issues/1Full protocol includes comparison of template to itself2018-02-15T15:24:59ZVedrana KRIVOKUCAFull protocol includes comparison of template to itselfThe "Full" protocol contains match scores of 0.5, which comes from comparing each template to itself. For my work, I have removed these scores manually, but I think it would be better to modify the protocol so that it only contains matc...The "Full" protocol contains match scores of 0.5, which comes from comparing each template to itself. For my work, I have removed these scores manually, but I think it would be better to modify the protocol so that it only contains match scores for DIFFERENT samples (i.e., not for a sample compared with itself). Below is an example of the EER results obtained for our 3 fingervein extractors (WLD, RLT and MC) with the Full protocol as it is VERSUS when the scores of 0.5 are removed:
WLD: 5.795% (current Full protocol) versus 9.545% (scores of 0.5 removed)
RLT: 14.545% (current Full protocol) versus 24.318% (scores of 0.5 removed)
MC: 2.477% (current Full protocol) versus 4.318% (scores of 0.5 removed)Vedrana KRIVOKUCAVedrana KRIVOKUCAhttps://gitlab.idiap.ch/bob/bob.kaldi/-/issues/6Results in the doc2017-10-13T12:29:25ZSébastien MARCELResults in the docCould you please augment the document with results and plots of experiments produced with bob.kaldi ?
eg. https://www.idiap.ch/software/bob/docs/bob/bob.bio.face/stable/baselines.html#baseline-results
cc @alexandre.nanchenCould you please augment the document with results and plots of experiments produced with bob.kaldi ?
eg. https://www.idiap.ch/software/bob/docs/bob/bob.bio.face/stable/baselines.html#baseline-results
cc @alexandre.nanchenMilos CERNAKMilos CERNAK2017-10-31https://gitlab.idiap.ch/bob/bob.ip.base/-/issues/6floating point exception (core dump) in DCTFeatures when block_size and block...2018-08-18T15:19:01ZAmir MOHAMMADIfloating point exception (core dump) in DCTFeatures when block_size and block_overlap are equalTo reproduce:
```python
>>> from bob.ip.base import DCTFeatures
>>> extractor = DCTFeatures(10, (14,14), (14,14), False, False)
>>> from bob.io.base.test_utils import datafile
>>> import bob.io.image
>>> import bob.io.base
>>> path = dat...To reproduce:
```python
>>> from bob.ip.base import DCTFeatures
>>> extractor = DCTFeatures(10, (14,14), (14,14), False, False)
>>> from bob.io.base.test_utils import datafile
>>> import bob.io.image
>>> import bob.io.base
>>> path = datafile('grace_hopper.png', 'bob.io.image')
>>> img = bob.io.base.load(path)
>>> import bob.ip.color
>>> gray = bob.ip.color.rgb_to_gray(img)
>>> gray.shape
(600, 512)
>>> dct = extractor(gray)
```https://gitlab.idiap.ch/bob/bob/-/issues/242Pip installations of Bob packages installs each package several times2017-08-07T12:16:52ZAmir MOHAMMADIPip installations of Bob packages installs each package several timesIn a clean environment without any bob package installed the `pip install bob` command will install each package several times which makes it impractical to install.
While this was done to circumvent the problem described here: https://s...In a clean environment without any bob package installed the `pip install bob` command will install each package several times which makes it impractical to install.
While this was done to circumvent the problem described here: https://stackoverflow.com/questions/14665936/import-egg-functionality-during-setup-in-a-buildout
I think this approach should have been avoided. a great example is Scipy which depends on Numpy during build. Looking at Scipy's `setup.py` I see that they have managed this pretty well: https://github.com/scipy/scipy/blob/master/setup.pyhttps://gitlab.idiap.ch/bob/bob.kaldi/-/issues/2Kaldi metadata2018-06-05T07:49:56ZMilos CERNAKKaldi metadataKaldi requires metadata to run advanced algorithms. The metadata consists of several text file lists, such as the list of train/test/dev utterances (sorted key utterance pairs), speaker to utterance file spk2utt (speaker utterances lists...Kaldi requires metadata to run advanced algorithms. The metadata consists of several text file lists, such as the list of train/test/dev utterances (sorted key utterance pairs), speaker to utterance file spk2utt (speaker utterances lists) and utterance to speaker file utt2spk (utterance speakers lists).
The idea is to add this metadata to prospective databases that could be used with Kaldi, such as AMI, NIST and so far.https://gitlab.idiap.ch/bob/bob.bio.face/-/issues/20Using super() for base class function calls2017-10-20T02:30:32ZManuel Günthersiebenkopf@googlemail.comUsing super() for base class function callsAfter bob.bio.base#64 is merged, we can go ahead and use super() to call base class functionality here, too. This mainly should affect constructor calls.After bob.bio.base#64 is merged, we can go ahead and use super() to call base class functionality here, too. This mainly should affect constructor calls.May 2017 Hackathonhttps://gitlab.idiap.ch/bob/bob.fusion.base/-/issues/2The HTER_w starts from 0%2019-02-26T03:07:52ZAmir MOHAMMADIThe HTER_w starts from 0%This is a bug in antispoofing.evaluation.This is a bug in antispoofing.evaluation.https://gitlab.idiap.ch/bob/bob.db.arface/-/issues/2Creating the database fails2018-07-10T19:03:33ZManuel Günthersiebenkopf@googlemail.comCreating the database failsWhen I try to run `bin/bob_dbmanage.py arface create`, I get the following error:
```
bob/db/arface/models.py", line 89, in __init__
bob.db.base.File.__init__(self, file_id = image_name, client_id = image_name[:5], path = image_name...When I try to run `bin/bob_dbmanage.py arface create`, I get the following error:
```
bob/db/arface/models.py", line 89, in __init__
bob.db.base.File.__init__(self, file_id = image_name, client_id = image_name[:5], path = image_name)
TypeError: __init__() got an unexpected keyword argument 'client_id'
```
I assume that there was an error while switching form `bob.db.verification.utils.File`, which accepts the `client_id` keyword to `bob.db.base.File`, which does not. While just using the database with the stored `db.sql3` database file, this works fine -- since the `bob.db.arface.File` constructor is never called. When trying to use the `File` class (as done during `create`), this issue shows up.
I assume that this issue shows up in many more databases.May 2017 HackathonAlain KOMATYAlain KOMATYhttps://gitlab.idiap.ch/bob/bob.fusion.base/-/issues/1push new things here2019-02-26T03:07:52ZAmir MOHAMMADIpush new things here@pkorshunov please open a pull request containing your new changes.@pkorshunov please open a pull request containing your new changes.Pavel KORSHUNOVPavel KORSHUNOVhttps://gitlab.idiap.ch/bob/bob.bio.vein/-/issues/10Implement dispersion plot for ROI metrics2020-11-17T17:57:18ZAndré AnjosImplement dispersion plot for ROI metricsIt would be nice to integrate a dispersion plot in the script `compare_rois.py`, which prints statistics for our established metrics.
The dispersion plot would plot changes in the metrics against changes in the scores. This will provide...It would be nice to integrate a dispersion plot in the script `compare_rois.py`, which prints statistics for our established metrics.
The dispersion plot would plot changes in the metrics against changes in the scores. This will provide a graphical visualization of the relationship between changes in the ROI metrics to changes in the scores.https://gitlab.idiap.ch/bob/annotation.video/-/issues/1This package needs to be integrated into bob/bob.ip.annotator2018-01-26T12:48:51ZAndré AnjosThis package needs to be integrated into bob/bob.ip.annotatorAs soon as it is done, we need to archive it.As soon as it is done, we need to archive it.Tiago de Freitas PereiraTiago de Freitas Pereirahttps://gitlab.idiap.ch/bob/bob.db.replaymobile/-/issues/3AttributeError: type object 'Attack' has no attribute 'sample_device_choices'2018-09-12T22:01:18ZAmir MOHAMMADIAttributeError: type object 'Attack' has no attribute 'sample_device_choices'```
db = bob.db.replaymobile.Database()
db.attack_sampling_devices()
---------------------------------------------------------------------------
AttributeError Traceback (most recent call last)
<ipython-input-8...```
db = bob.db.replaymobile.Database()
db.attack_sampling_devices()
---------------------------------------------------------------------------
AttributeError Traceback (most recent call last)
<ipython-input-8-9be7b1aeccd8> in <module>()
----> 1 db.attack_sampling_devices()
/home/amir/idiap/git/bobs/bob.bio.face/src/bob.db.replaymobile/bob/db/replaymobile/query.py in attack_sampling_devices(self)
304 """Returns sampling devices available in the database"""
305
--> 306 return Attack.sample_device_choices
307
308 def attack_sample_types(self):
AttributeError: type object 'Attack' has no attribute 'sample_device_choices'
```Sushil BHATTACHARJEESushil BHATTACHARJEEhttps://gitlab.idiap.ch/bob/bob.extension/-/issues/28Automatic docstring generator should adopt one of "Napoleon" styles2019-09-04T13:12:40ZAndré AnjosAutomatic docstring generator should adopt one of "Napoleon" stylesWhen we wrote our documentation generator in [documentation.h](bob/extension/include/bob.extension/documentation.h), there was not a set standard for documenting functions in sphinx. Since release 1.3 (which is now required by most of ou...When we wrote our documentation generator in [documentation.h](bob/extension/include/bob.extension/documentation.h), there was not a set standard for documenting functions in sphinx. Since release 1.3 (which is now required by most of our software), there is:
http://www.sphinx-doc.org/en/stable/ext/napoleon.html
The napoleon extension is already loaded by most of our software, so it makes sense that the generator follows one of the docstring standards defined on that manual page. Some of our python-based docstrings already follow the "Google" style:
http://www.sphinx-doc.org/en/stable/ext/example_google.html#example-google
It would be good to get the C++ docs matching that.May 2017 HackathonPavel KORSHUNOVPavel KORSHUNOVhttps://gitlab.idiap.ch/bob/bob.ap/-/issues/3C++ code documentation is not using the standard2017-08-17T14:18:22ZManuel Günthersiebenkopf@googlemail.comC++ code documentation is not using the standardIn most other packages containing C++ code, we have switched to the ``bob::extension::`` documentation classes. These classes have the advantage that:
1. We get a unified documentation.
2. We make sure all parameters are documented, ...In most other packages containing C++ code, we have switched to the ``bob::extension::`` documentation classes. These classes have the advantage that:
1. We get a unified documentation.
2. We make sure all parameters are documented, including the expected type
3. We assure that the parameter names in the documentation and what the function expects are identical.
Additionally, we normally use the ``BOB_TRY`` and ``BOB_CATCH_...`` macros in order to automatically catch any exception and provide a nice Python exception message, and make the code more readable.
This package does use neither of them.https://gitlab.idiap.ch/bob/bob.ip.optflow.liu/-/issues/1Segmentation fault ...2017-09-17T15:14:31ZAndré AnjosSegmentation fault ...*Created by: guillaume-heusch*
Hi,
I wrote a little test script using Liu's optical flow (using conjugate gradient), and I ran into a strange behaviour:
- when I load two color images of size 72*72 pixels, everything is fine :)
-...*Created by: guillaume-heusch*
Hi,
I wrote a little test script using Liu's optical flow (using conjugate gradient), and I ran into a strange behaviour:
- when I load two color images of size 72*72 pixels, everything is fine :)
- when I load the same color images, downscaled to 36*36 pixels (downscaling done by bob), I get a segmentation fault.
Note that the two sets of two images have been processed the exact same way after having been loaded and before the call to bob.ip.optflow.liu.cg.flow: casted to float and divided by 255, and then converted to grayscale ...
Thanks,
Guillaume
Guillaume HEUSCHGuillaume HEUSCHhttps://gitlab.idiap.ch/bob/bob.ip.color/-/issues/1python API crashing with C++ exceptions2017-10-21T18:59:38ZAndré Anjospython API crashing with C++ exceptions*Created by: siebenkopf*
The color conversion functions in the Python API, such as ``bob.ip.color.rgb_to_gray`` might crash with a C++ exception, which is not caught correctly inside the python bindings.
Solution: go through the func...*Created by: siebenkopf*
The color conversion functions in the Python API, such as ``bob.ip.color.rgb_to_gray`` might crash with a C++ exception, which is not caught correctly inside the python bindings.
Solution: go through the functions and add a ``BOB_TRY ... BOB_CATCH_FUNCTION`` wrapper around them (see http://pythonhosted.org/bob.extension/cpp_api.html#c.BOB_CATCH_FUNCTION).