A frozen pre-trained LLM layer inserted between a CNN encoder and decoder improves medical image segmentation across ultrasound, dermoscopy, polyp, and CT benchmarks with few added trainable parameters.
discriminability: Batch spectral penalization for adversarial domain adaptation
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
-
Pre-Trained LLM is a Semantic-Aware and Generalizable Segmentation Booster
A frozen pre-trained LLM layer inserted between a CNN encoder and decoder improves medical image segmentation across ultrasound, dermoscopy, polyp, and CT benchmarks with few added trainable parameters.