Pain Expression Recognition Based on pLSA Model

Figure 1

The processing pipeline of the “bag-of-words” representation: (a) give an image, (b) divide into 𝐿×𝐿 blocks, (c) represent each block by a “visual word,” and (d) ignore the ordering of words and represent the facial image as a histogram over “visual words.”

(a)

(b)

(c)

(d)