Table 7 Number of extracted keyframes in the implemented experiments
From: Bilingual video captioning model for enhanced video retrieval
Criteria | Training | Validation | Testing | Total |
|---|---|---|---|---|
First experiment | ||||
 Time-based 50 | 21,248 | 1843 | 13,826 | 36,917 |
 Similarity-based 90 | 19,499 | 1744 | 12,866 | 34,109 |
Second experiment | ||||
 Time-based 25 | 43,097 | 3730 | 28,008 | 74,835 |
 Similarity-based 90 | 37,374 | 3367 | 24,669 | 65,410 |
Third experiment | ||||
 Time-based 25 | 43,097 | 3730 | 28,008 | 74,835 |
 Similarity-based 95 | 41,128 | 3597 | 27,013 | 71,738 |
Original experiment | 304,564 | 26,550 | 210,752 | 541,866 |