Skip to main content
. 2014 Feb 7;35(8):4219–4235. doi: 10.1002/hbm.22472

Figure 3.

Figure 3

Feature extraction. For each vowel, v, a patch I(v) is used to construct the features. f 1 gives the intensity variation inside I(v) and f 2…, f 16 represent the textural information. The final feature vector is ϕ(v). By its contribution ϕ(v) is 16 times the number of voxels in I(v). Note that I(v) is three‐dimensional. [Color figure can be viewed in the online issue, which is available at http://wileyonlinelibrary.com.]