D-MLM models medical image and text features as Gaussian distributions and uses LLM-structured reports to achieve state-of-the-art results on several chest X-ray benchmarks.
In: European Conference on Computer Vision
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
-
Distribution-Based Masked Medical Vision-Language Model Using Structured Reports
D-MLM models medical image and text features as Gaussian distributions and uses LLM-structured reports to achieve state-of-the-art results on several chest X-ray benchmarks.