the key method. Takes image on input, vector of ERStat is output for the first stage, input/output for the second one.

! Create an Extremal Region Filter for the 1st stage classifier of N&M algorithm Neumann L., Matas J.: Real-Time Scene Text Localization and Recog

Fit line from two points out a0 is the intercept out a1 is the slope

Create a dummy classifier that accepts all regions */ Ptr<ERFilter::Callback> loadDummyClassifier();

add history chunk to a connected component

Evaluates if a pair of regions is valid or not using thresholds learned on training (defined above)

Evaluates if a set of 3 regions is valid or not using thresholds learned on training (defined above)

Compare doubles by relative error.

Interface functions

Evaluates if a set of more than 3 regions is valid or not using thresholds learned on training (defined above)

construct and solve various formulations

Stratified cross validation

label: label name, start: begin of each class, count: #data of classes, perm: indices to the original data perm, length l, must be allocated before ca

! * MSERsToERStats function converts MSER contours (vector<Point>) to ERStat regions. * It takes as input the contours provided by the OpenCV MSER

Computes -log10(NFA). NFA stands for Number of False Alarms: */

changed bu liuruoze

Deletes a tree of ERStat regions starting at root. Used only internally to this implementation.