Commit Graph

153 Commits

Author SHA1 Message Date
Davis King e00f6f63ca Replaced the raw_input() calls with a call into dlib since raw_input() isn't
available in python3.
2015-03-22 18:45:08 -04:00
Davis King 4bbfe512ae Made the PYTHON3 cmake switch a proper cmake option. 2015-03-07 13:29:23 -05:00
Davis King c335bf6756 Fixed the python doc string so that it's correct relative to the python
API.  Also made find_candidate_object_locations() correctly include any
input boxes in the call to the C++ version of the function.
2015-01-03 23:26:28 -05:00
Vinh Khuc 173bb97f33 Explained the difference between pykvals in the Python API and kvals in C++ code 2015-01-03 15:03:05 -05:00
Vinh Khuc bf06ce7185 Added Python-binding for find_candidate_object_locations 2015-01-03 01:11:52 -05:00
Davis King 1ab3482597 Clarified a few comments and simplified the serialization code a bit.
Also just cleaned up a few minor details.
2014-12-27 15:30:56 -05:00
Patrick Snape 37af35b55e Re-add the cached object detector
A little bit hacky, but should be fine. Supports both fhog
detectors and the "cached" simple_object_detector. Also, maintains
the upscale parameter for testing
2014-12-12 16:22:57 +00:00
Patrick Snape 30869fbe03 Change logic for upsampling printing 2014-12-11 15:00:39 +00:00
Patrick Snape dd19ce846e Update the interface to be more Pythonic
This is the biggest change so far. Now, there are two different
classes of interface. One where you pass ONLY file paths,
and one where you pass ONLY Python objects.

The file paths are maintained to keep a matching interface with
the C++ examples of dlib. So shape predicition and object
detection can be trained using the dlib XML file paths and then
serialize the detectors to disk.

Shape prediction and object detection can also be trained using
numpy arrays and in-memory objects. In this case, the predictor
and detector objects are returned from the training functions.
To facilitate serializing these objects, they now have a 'save'
method.

Tetsing follows a similar pattern, in that it can take either XML
files are or in-memory objects. I also added back the concept of
upsampling during testing to make amends for removing the
simple_object_detector_py struct.
2014-12-11 14:06:05 +00:00
Patrick Snape 8db3f4e572 Add a save method to detectors and predictors
Also, removed the saving of the upsample which I missed from
before (since I'm not using the struct now). I understand why
the upsample was being saved, but I don't necessarily agree it
is particularly useful as you should really be upsampling on
a case by case basis at test time.
2014-12-11 12:34:20 +00:00
Patrick Snape 5b485a628b Properly handle turning the GUI off 2014-12-11 12:05:10 +00:00
Patrick Snape e801bd6ab5 Refactor the GUI code out
I also cleaned up a bunch of code. I'm not sure why the
simple_object_detector was keeping track of the upsample amount,
since it can't even be passed as an argument to the constructor.
Therefore, I removed the simple_object_detector_py and the second
declaration of the hog object detector. I also changed the
view code to optionally take keyword args of color and added
a single view of a rectangle.

Finally, I added viewing of the shape parts.
2014-12-11 11:46:01 +00:00
Patrick Snape 697aecb420 Add rendering of faces (lines)
Can either be a list of full_object_detections or a single
full_object_detection. I couldn't get the vector type to work
for full_object_detection due to a template error.
2014-12-11 10:28:14 +00:00
Patrick Snape 32ad0ffaef Make shape predictor pickleable
Fix typo as well
2014-12-10 18:07:38 +00:00
Patrick Snape e3aee32f34 Add wrappers for the shape predictors
This includes the full_object_detection, a new struct in the same
vein as the simple_object_detector_training_options and of
course, the shape predictor classes themselves.

All of training, fitting and testing are wrapped.
2014-12-10 17:45:51 +00:00
Patrick Snape 315a2b1cb1 Add a new conversion header
This deals with converting python objects to dlib objects
2014-12-10 17:45:51 +00:00
Patrick Snape 68ae858f27 Refactor rgb_pixel out of object detection
Also, move the vectorize template into its own header to
stop having to declare it again in vector.
2014-12-10 17:45:51 +00:00
Patrick Snape c0d0adba13 Add a list of points
Also, change point x and y to be properties
2014-12-10 17:45:51 +00:00
Patrick Snape 85f0c0ffdc Wrap the dlib point for Python 2014-12-10 17:45:51 +00:00
Patrick Snape 70db61c5fb First attempt at adding to the boost API
Exposes the test method, but takes in images and bounding boxes.
Seems fairly simply to extend the API
2014-12-10 17:45:51 +00:00
Davis King cdbc1919d4 more cmake changes to avoid cmake warnings 2014-12-06 08:38:04 -05:00
Davis King db3eaa728b Changed python cmake scrips to have an option to compile Python 3 libraries. 2014-11-26 16:28:52 -05:00
Davis King cc70a8e0fb Merged in Patrick Snape's patch that allows the python bindings to
be built without GUI support.
2014-11-15 09:28:48 -05:00
Davis King c9f24f3da4 upped the imglab version number 2014-08-21 21:27:51 -04:00
Davis King cd71dab3f2 Updated the Python API, train_simple_object_detector() so you can call it
directly on already loaded data rather than needing to use an XML file as
input.
2014-08-12 19:47:41 -04:00
Davis King a6dc296219 Fixed a bug where the imglab tool would sometimes start out with an annoyingly
small window when the first image was bigger than the screen.
2014-07-20 14:47:05 -04:00
Davis King c423283c05 Added set_prior() to the python interfaces of the relevant trainer objects. 2014-05-24 10:18:31 -04:00
Davis King 312157ab18 Added epsilon to the python object detection training API. 2014-03-09 12:10:15 -04:00
Davis King e79a764834 removed cruft 2014-03-08 12:14:24 -05:00
Davis King 2369ae1e57 Made the python object detection tools process color images
in color instead of always converting to grayscale.
2014-03-02 11:19:06 -05:00
Davis King a612c0317d Improved error message. 2014-03-02 11:08:20 -05:00
Davis King 75fbc56c92 Fixed name conflicts that prevented the python bindings from compiling in visual sudio 2013. 2014-03-01 21:42:01 -05:00
Davis King 4a9be7bba5 improved python docs 2014-02-23 16:18:04 -05:00
Davis King 2293c91e00 Cleaned up the new python object detection interface a little. 2014-02-23 16:02:52 -05:00
Davis King 8f6b76e7ca Gave the image_window the ability to display fhog object detectors. 2014-02-23 15:24:22 -05:00
Davis King ddc44067b4 Added a simple python interface for training fhog object detectors. 2014-02-23 14:39:59 -05:00
Davis King 29658c4c76 Minor CLI cleanup 2014-02-22 22:03:49 -05:00
Davis King 89bc542712 fixed grammar in comment 2014-02-22 22:00:07 -05:00
Davis King 017fedcc41 updated about 2014-02-11 20:23:21 -05:00
Davis King 01eaaa563a Made imglab halt at a convenient place when you are using ctrl+key up/down to
auto label boxes.  It will also auto-skip empty images when you hold ctrl.
2014-02-11 20:22:15 -05:00
Davis King aa6919d304 Made --stats print out info on unlabeled boxes. 2014-02-11 20:19:57 -05:00
Davis King a20f401eec Gave imglab the ability to propagate rectangle labels between images. 2014-02-11 19:03:08 -05:00
Davis King 374d5253f8 Updated version number and about string to say that you can now ctrl+left click
to change the label for a rectangle.
2014-02-10 20:40:24 -05:00
Davis King a20f154483 Added a python interface to the fhog_object_detector and related tools. 2014-02-09 12:56:35 -05:00
Davis King ef6a3a05a6 Added a readme file to imglab that gives a little introduction 2014-01-09 22:15:17 -05:00
Davis King 1cf740156d Added --flip to imglab 2013-11-29 22:35:52 -05:00
Davis King 48413ae845 Updated about window text for imglab. 2013-11-22 18:45:06 -05:00
Davis King 647285e808 Updated imglab so it loads and stores the new ignore option from the gui
to/from the xml files.
2013-11-21 20:40:12 -05:00
Davis King 5edb5662e0 Added the --add option to the imglab tool. 2013-11-16 16:28:13 -05:00
Davis King 309902b81a clarified specs 2013-09-21 15:18:29 -04:00