Automatic Representation and Segmentation of Video Sequences via a Novel Framework Based on the D-EVM and Kohonen Networks

Table 8

Size of the 3D-EVMs generated for clusters associated with Billiard Game Video Sequence under a SOM with 15 neurons and removal of temporal component.

Class	Members	EVM size
0	244,714	40,280
1	187,126	22,722
2	2,048	930
3	201,427	30,670
4	3,428	2,062
5	210,437	42,532
6	189,185	25,880
7	8,776	3,966
8	11,881	5,808
9	548,273	10,294
10	143,238	25,930
11	29,990	8,946
12	16,003	7,350
13	4,944	2,732
14	3,522	1,558
Total	1,804,992	231,660