Multilabel Image Annotation Based on Double-Layer PLSA Model

Figure 3

The “image visual-word” matrix of the image low-level feature.