A weakly supervised bag-of-visual-words pipeline over frozen pathology embeddings produces interpretable lung adenocarcinoma pattern maps and matches or beats a mean-pooling SVM baseline on tumour detection and grade classification.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2026 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Bag-of-Visual-Words for Spatial Mapping of Lung Adenocarcinoma Growth Patterns
A weakly supervised bag-of-visual-words pipeline over frozen pathology embeddings produces interpretable lung adenocarcinoma pattern maps and matches or beats a mean-pooling SVM baseline on tumour detection and grade classification.