OrgLion-ML
/
dlib
mirror of https://github.com/davisking/dlib.git
1
0
Fork 0
Code Issues Releases Wiki Activity

improved comments

This commit is contained in:
Davis King 2012-03-22 18:44:47 -04:00
parent ae8e1b5f7d
commit 1d3b429f40
1 changed files with 12 additions and 13 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

25
examples/train_object_detector.cpp
View File

@ -2,7 +2,7 @@
/*
This is an example showing how you might use dlib to create a reasonably
function command line tool for object detection. This example assumes
functional command line tool for object detection. This example assumes
you are familiar with the contents of at least the following example
programs:
- object_detector_ex.cpp
@ -13,17 +13,15 @@
This program is a command line tool for learning to detect objects in images.
Therefore, to create an object detector it requires a set of annotated training
images. To create this annotated data you will need to compile the imglab tool
included with dlib. To do this, go to the tools/imglab folder and type the
following:
images. To create this annotated data you will need to use the imglab tool
included with dlib. It is located in the tools/imglab folder and can be compiled
using the following commands.
cd tools/imglab
mkdir build
cd build
cmake ..
make
Note that you may need to install CMake (www.cmake.org) for this to work. Also,
if you are using visual studio then you should use the following command to
compile instead of "make"
cmake --build . --config Release
Note that you may need to install CMake (www.cmake.org) for this to work.
Next, lets assume you have a folder of images called /tmp/images. These images
should contain examples of the objects you want to learn to detect. You will
@ -39,13 +37,14 @@
with boxes. Note that it is important to label all the objects since any object
not labeled is implicitly assumed to be not an object we should detect.
Once you finish labeling objects go to the file menu, click save, and close the
program. This will save the object boxes back to mydataset.xml. You can verify
Once you finish labeling objects go to the file menu, click save, and then close
the program. This will save the object boxes back to mydataset.xml. You can verify
this by opening the tool again with
./imglab mydataset.xml
and observing that the boxes are present.
Returning to the present example program, we can now issue the command
Returning to the present example program, we can compile it using cmake just as we
did with the imglab tool. Once compiled, we can issue the command
./train_object_detector -tv mydataset.xml
which will train an object detection model based on our labeled data. The model
will be saved to the file object_detector.svm. Once this has finished we can use
@ -55,8 +54,8 @@
be indicated by a red box.
There are also a number of other useful command line options in the current
example program which you can explore below.
There are a number of other useful command line options in the current example
program which you can explore below.
*/