We test four conditions that can be an indication of a pitching scene, and four boolean variables C A, C L, C V, and C P are set as follows: i) Area ratio condition ( C A ): Area ratio [36], R a, is computed from the binary image I f.
ii) Layout condition ( C L ): In pitching scenes, the lower half of the image usually contains more baseball field pixels [36].
(11) In pitching scenes, a strong valley is usually found in the left hand side of this distribution [36], as shown in Figure 12c, due to the presence of the pitcher.
When he pitched, the scene was as electric as any he had experienced in the major leagues.
Jean was a fan of the series and pitched the scene.
The false detections indicate that certain scenes are wrongly identified as the pitching view frames so the reverse search stops before reaching the targeted ones.
If 25%≤R a ≤45% then C A =1; else 0. Figure 12 Pitching scene detection.
Figure 12b shows an example binary image from a pitching scene.
(a) a sample pitching scene, (b) detected field pixels (shown in black), (c) vertical profile V(i) of the field pixels, and (d) 16 blocks dividing the image.
If the image intensity of these regions is greater than the average intensity of the image, C P =1; else 0. From our observations, almost all pitching scenes fulfill the condition C V, but not C A. Thus, unlike in [15], we declare the i th frame as a pitching scene if the following expression yields TRUE: C pitch = C V · ( C L + C A + C P ).
And then she would pitch another scene to counter that scene.
