Figure 13From: Contextual Classification of Image Patches with Latent Aspect ModelsImage patch classification examples at TEER. Results provided by the following: first column, K-means empirical distribution; second column, aspect model 1; third column, aspect model 2. The total number of correctly classified patches (man-made + natural) is given per image. The five rows illustrate cases where aspect model 2 outperforms the other approaches. In the fifth row, an extreme example of a strong natural context that is correctly identified by aspect model 2 leads to the classification of all regions as natural (though some should be labeled as man-made).Back to article page