Real-Time Vehicle Detection Using Cross-Correlation and 2D-DWT for Feature Extraction

Algorithm 1

 Input: is a set of labeled examples where belongs to , .

 Initialization: for .

 For :

 Train the weak learner based on distribution D_t

 Obtain the weak hypotheses h_t:

 Select h_t with low weighted error:

 If , then set and abort loop

 Choose

 Update :

 Where is a factor of normalization (chosen in a way that is a distribution)

 The final hypothesis is given as