Skip to main content
. 2012 Jul 13;6:45. doi: 10.3389/fncom.2012.00045

Figure 4.

Figure 4

The 10 shape prototypes used in conjunction with CT to encode the ROIs comprising the scenes (see section 7.1). Each ROI detected in a scene was represented by a 10-dimensional vector of its respective similarities to these 10 images.