LEAP:D swaps manual prompts for 8 learnable tokens in a CLIP-guided Faster R-CNN, improving mAP50 from 39.7 to 42.1 on VisDrone validation, with no cross-domain evaluation.
Our comparative experiments revealed that LEAP:D outperforms baseline models and other state-of-the-art methods
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2024 1verdicts
REJECT 1representative citing papers
citing papers explorer
-
LEAP:D -- A Novel Prompt-based Approach for Domain-Generalized Aerial Object Detection
LEAP:D swaps manual prompts for 8 learnable tokens in a CLIP-guided Faster R-CNN, improving mAP50 from 39.7 to 42.1 on VisDrone validation, with no cross-domain evaluation.