IMCBench is a new benchmark for image-grounded multi-turn medical conversations that evaluates eight multimodal LLMs on safety, accuracy, and uncertainty, finding Claude Opus highest overall but safety drops for malignant and rare conditions.
hub
The HAM10000 dataset, a large collection of multi -source dermatoscopic images of common pigmented skin lesions
15 Pith papers cite this work, alongside 3,113 external citations. Polarity classification is still indexing.
15
Pith papers citing it
3,113
external citations · Crossref
hub tools
citation-role summary
background 1
citation-polarity summary
roles
background 1polarities
background 1representative citing papers
DermAgent orchestrates seven vision-language tools in a Plan-Execute-Reflect loop with dual-modality retrieval from 413k cases and a critic module to outperform GPT-4o by 17.6% in zero-shot dermatological diagnosis accuracy.
MedCore achieves 60% parameter and 58.4% FLOP reduction on MedSAM with Dice 0.9549 and preserved boundary metrics via dual-intervention pruning and a new boundary leverage principle.
Rough-set analysis finds 16.4% of 305 concept profiles in Derm7pt inconsistent (306 images), capping hard CBM accuracy at 92.1%; symmetric filtering produces a 705-image consistent benchmark where EfficientNet-B5 reaches 0.90 label accuracy.
A CNN-based discrete diffusion method refines sparse contours from segmentation masks using simplified denoising steps and minimal post-processing, outperforming baselines on small medical and environmental datasets while running 3.5 times faster.
The authors introduce predicted-weighted balanced accuracy (pBA), a utility-weighted evaluation metric that uses predicted subconcept posteriors to reduce bias from within-class heterogeneity in imbalanced data.
FSS-TIs models cross-domain few-shot segmentation as an ODE process with Fourier-based spectral perturbations to create domain-agnostic features and enable effective fine-tuning on limited support samples.
MARVEL introduces a multi-expert NvMF-based system with an outlier expert that reduces FPR95 in OOD detection on medical datasets by 8-37%.
FedSSG generates and shares synthetic samples within a federated setup to reduce class imbalance and domain shift problems in medical image classification.
IViT applies quadratic programming to a pre-trained Vision Transformer with a multi-objective loss, achieving 93.80% accuracy on six skin disease datasets (0.21% below baseline) while reducing feature redundancy by 29.5% and producing clinically consistent activations.
Cascade classification improves macro F1 over single-stage for some models by allowing sensitivity control but reveals a large generalization gap on external clinical data.
Describes a methodology and the resulting dataset of 1,026 dermoscopic images with structured metadata and verified diagnostic labels for medical informatics research.
MedGemma 1.5 4B reports absolute gains of 11% on 3D MRI classification, 3% on 3D CT, 47% macro F1 on pathology slides, 35% IoU on anatomical localization, and 5-22% on clinical QA tasks over MedGemma 1.
Fine-tuned MedGemma outperforms untuned GPT-4 in zero-shot medical image disease classification, achieving 80.37% versus 69.58% mean test accuracy with higher sensitivity for cancer and pneumonia.
Post-hoc normalizing flows for OOD detection in medical imaging achieve 84.61% AUROC on MedOOD and 93.8% on MedMNIST, outperforming ViM, MDS, and ReAct.
citing papers explorer
-
Concept Inconsistency in Dermoscopic Concept Bottleneck Models: A Rough-Set Analysis of the Derm7pt Dataset
Rough-set analysis finds 16.4% of 305 concept profiles in Derm7pt inconsistent (306 images), capping hard CBM accuracy at 92.1%; symmetric filtering produces a 705-image consistent benchmark where EfficientNet-B5 reaches 0.90 label accuracy.