Information Technology Reference
In-Depth Information
Fig. 16 Plot of the recognition results obtained by the HPaSMM method with the LOOCV
procedure. Ground truth is drawn in red while the recognition results are plotted in blue.
Detected referees whistles are indicated by dotted lines. Numbers on the vertical axis corre-
spond to activity phases numbering as described in Section 4.2. The horizontal axis denotes
the numbering of the successive group of images (with parameter k group value set to 8).
Ta b l e 1 Correct recognition rates obtained by the HPaSMM method for the eight activity
phases. Experiments were conducted using the LOOCV test method, without considering the
events “7-meter throw” and “Jump ball at the court center”. The number of images corre-
sponding to each activity phase is also indicated.
Recognition
#
rate
images
Slowly going into attack
68.1
728
Attack against set-up defense
92.4
3776
Offense free-throw or timeout
87.1
1056
counterattack, fast break
95.6
1280
Returning, preventing from fast break
88.6
280
Slowly returning in defense
93.6
1112
Defense
100
3592
Defense free-throw or timeout
81
928
Comparison with the HPaHMM method using audio information and LOOCV
Referees whistles were integrated in the HPaHMM method in a similar way than
with the HPaSMM one. Hence, each segment is decoded separately and the LOOCV
test method defined above is also considered with HPaHMM method.
We report here the result of the HPaHMM method when processing the set of
trajectories of about 8 minutes and 30 seconds (still omitting the “7-meter throw”
 
Search WWH ::




Custom Search