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

Updated example program to work best with the new code

This commit is contained in:
Davis King 2013-08-06 00:33:50 -04:00
parent 4b29dec604
commit 36d5677a26
1 changed files with 6 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

7
examples/object_detector_ex.cpp
View File

@ -173,9 +173,14 @@ int main()
// The hashed_feature_image in the scanner needs to be supplied with a hash function capable // The hashed_feature_image in the scanner needs to be supplied with a hash function capable
// of hashing the outputs of the hog_image. Calling this function will set it up for us. The // of hashing the outputs of the hog_image. Calling this function will set it up for us. The
// 10 here indicates that it will hash hog vectors into the range [0, pow(2,10)). Therefore, // 10 here indicates that it will hash HOG vectors into the range [0, pow(2,10)). Therefore,
// the feature vectors output by the hashed_feature_image will have dimension pow(2,10). // the feature vectors output by the hashed_feature_image will have dimension pow(2,10).
setup_hashed_features(scanner, images, 10); setup_hashed_features(scanner, images, 10);
// We should also tell the scanner to use the uniform feature weighting scheme
// since it works best on the data in this example. If you don't call this
// function then it will use a slightly different weighting scheme which can give
// improved results on many normal image types.
use_uniform_feature_weights(scanner);
// We also need to setup the detection templates the scanner will use. It is important that // We also need to setup the detection templates the scanner will use. It is important that
// we add detection templates which are capable of matching all the output boxes we want to learn. // we add detection templates which are capable of matching all the output boxes we want to learn.