-
Simplify SingleBuffer since controlling the total amount of frames being processed works quite well. Add a double-buffer scheme, which leads to a really marginal performance improvement (but I think it's cool).
-
Rather than putting limits on the number of items in the buffers of individual processing stages of a stream, put a limit on the total number of frames actively being processed by the stream.
-
Add Show2Transform (should replace ShowTransform eventually), which displays input images using QT, rather than going through openCV. Additional modifications to support doing this (since qt is quite particular about how GUI elements are created/modified). Create a QApplication instead of a QCoreApplication (if not building BR_EMBEDDED), restructure br.cpp's main to run its main work in a separate thread while the main thread sits in an event loop.
-
Add StreamTransform which represents applying a series of transforms to some sequential input (e.g. a video). Handles reading video frames in and processing them as they become available (rather than loading the entire video into memory before processing it). The basic idea is a pipeline model, every child transform of StreamTransform is run by a separate thread, (as is reading in video frames). A thread waits for a frame as input, when input becomes available performs its operation, and then places the frame on the following stage's input buffer. Single threaded processing is not yet implemented. Moved CompositeTransform to openbr_plugin.h Add ContractTransform -- combines all input templates into a single template Preliminary handling of time varying transforms in DistributeTemplateTransform, the current behavior is not actually correct but will work for the single input case.
-
Make a call to qSort prior to shuffling the records received by a dbGallery this makes the output of a dbGallery deterministic for a given query (as long as the seed is set).
-
Handle enrollment of multiple independent templates by creating a single element template list from each input template, then processing these TemplateLists in parallel. This changes the basic operation for processing templates derived from a sinlge image from project(template, template) (a single input/single output function) to project(templateList, templateList), which allows for a transform to output a variable number of templates from a given input (e.g. a detector can output multiple templates, or a filter can remove templates that don't meet some criterion). The basic parallel execution code loop is implemented using DistributeTemplateTransform, instead of through Template::Project(list), and a DistributeTemplateTransform is automatically prepended to input enrollment algorithms. This avoids the problem of having to decide whether or not a split should take place on every call to project(templatelist), and since the distribute transform is just a normal transform, it can be applied more than once in a given algorithm (when this is desirable). Implement ExpandTransform as a basic Transform that take a template list as input, and expands it for its output (in project(templatelist)), due to the restructuring ExpandTransform no longer has to essentially mimic the functionality of PipeTransform to function correctly. This commit breaks binary compatibility (of saved models), so updates to the pre-trained face recognition and gender classification models are included.
-
getBool("X") --> get<bool>("X",false) and appendLandmarks --> appendPoints
