A training-free editing pipeline that uses a large emotion-annotated dataset and a fine-tuned CLIP model to modify only the parts of an image that convey a target emotion.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Moodifier: MLLM-Enhanced Emotion-Driven Image Editing
A training-free editing pipeline that uses a large emotion-annotated dataset and a fine-tuned CLIP model to modify only the parts of an image that convey a target emotion.