|
I am doing image classification. I am extracting local features from images. Each local feature depends on a parameter w. Finally the local features are histogrammed to get a final feature vector which represents the image. Based on these feature vectors (histograms) I am doing the final classification. I am emperically tuning the parameter w. Are there any way I can learn this parameter? For example: 1. some kind of enery minimization 2. Maximum likelihood 3. etc. |
