Locate before Answering: Answer Guided Question Localization for Video Question Answering
Video question answering (VideoQA) is an essential task in vision-language understanding, which has attracted numerous research attention recently. Nevertheless, existing works mostly achieve promising performances on short videos of duration within 15 seconds. For VideoQA on minute-level long-term...
Salvato in:
| Autori principali: | , , , , , |
|---|---|
| Natura: | Articolo |
| Pubblicazione: |
2022
|
| Soggetti: | |
| Accesso online: | https://arxiv.org/abs/2210.02081 http://arxiv.org/abs/2210.02081 |
| Tags: |
Nessun Tag, puoi essere il primo ad aggiungerne!!
|