Open embeh opened 8 years ago
HI, to avoid using any external toolbox, I just simply used the function 'extractHOGFeatures' from the Matlab computer vision system toolbox in this code.
You can just simply replace the feature extraction function using any other feature extraction method such as lbp, sift and hog from the open source vlfeat toolbox, by modifying the obtain_features.m file.
Running example_detection in Matlab 2015b, I get this error message:
Training the 1-th weak regressor ... Undefined function or variable 'extractHOGFeatures'.
Error in obtain_features (line 28) hogf = extractHOGFeatures(tmp);
Error in train_sdm (line 42) feature_matrix(:,flag) = obtain_features(train_img(i).data, current_shape(:,flag), cr_model);
Error in example_detection (line 95) cr_model.model = train_sdm(train_img, train_init_shape, train_gt_shape, cr_model);
Looks like this code requires the Computer Vision System Toolbox?