NIH ChestX-ray14
PulseAugur coverage of NIH ChestX-ray14 — every cluster mentioning NIH ChestX-ray14 across labs, papers, and developer communities, ranked by signal.
2 day(s) with sentiment data
-
New method improves chest X-ray disease localization using distilled evidence
Researchers have developed APED-CXR, a new method for improving the localization of diseases in chest X-rays. This approach distills spatial evidence from a multimodal teacher model into a smaller, image-only student mo…
-
Vision Transformer models exhibit annotation bias in medical imaging
Researchers have identified a significant bias in Vision Transformers (ViTs) used for medical imaging analysis. A specific attention head, L8H10 in ViT-Base 16, was found to focus on corner text annotations rather than …
-
New method quantifies uncertainty in black-box vision models
Researchers have developed SpatialUQ, a novel post-hoc uncertainty quantification method for black-box vision models, particularly in clinical settings. This method measures the divergence between a global prediction an…
-
New GRAPE architecture enhances medical image diagnosis with graph-augmented explanations
Researchers have developed GRAPE (Graph-Augmented Prototype Explanations), a novel architecture designed to improve medical image diagnosis systems. GRAPE addresses limitations of existing prototype-based classifiers by…
-
New method decomposes ML model interactions into uniqueness, redundancy, and synergy
Researchers have developed a new method called Stochastic Hi-Fi to better understand the interactions within machine learning models. This technique decomposes feature importance into uniqueness, redundancy, and synergy…