OSSL is the first self-hosted, mood-annotated video-music dataset, and a video adapter on MusicGen-Medium improves film music generation over text-only baselines.
To show the effectiveness of our dataset, we adapted a text-to-music generation model with video conditions and fine-tuned it on our dataset
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.SD 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Video-Guided Text-to-Music Generation Using Public Domain Movie Collections
OSSL is the first self-hosted, mood-annotated video-music dataset, and a video adapter on MusicGen-Medium improves film music generation over text-only baselines.