Skip to main content
. 2012 Oct 4;7(10):e46686. doi: 10.1371/journal.pone.0046686

Figure 8. One-vs.-One discriminative natural action structures.

Figure 8

(A), One-vs.-One discriminative NASs in three pairs of actions in the KTH dataset. The locations and NASs in each individual action are indicated by the same color. (B), Same format as (A) for the Weizmann dataset. Note that each NAS is a set of patch sequences and also that each NAS shown is the average of all patch sequences that share the same structural indices.