Adapted DINOv2 generalises better than task-specific medical segmentation nets
Only a small part of DINOv2 was adapted; in-distribution U-Nets stayed sharper and caught small lesions more often.
Stories per day · 30 days
1 RSS
Relateddiffusion3ct2echocardiography2landmark-detection2super-resolution2vascusr2agents1benchmark1cardiac-analysis1chest-x-ray1code-execution1contrastive-learning1dataset1diffpt1diffusion-models1dinov21domain-shift1em-detr1
37 stories, newest first · one card per story
Stories tagged #medical-imaging
EM-DETR foundation detector trained across heterogeneous medical datasets
Authors report about 1 AUC point higher on four CXR findings and 2–3 mAP50 points over single-dataset EM-DETR.
P2Echo segments multi-view echocardiography from text prompts with one model
A routing module switches decoder pathways by view and structure, matching specialists at lower compute.
LUCQUE releases annotated Québec LDCT lung-screening data for AI evaluation
The set has 1,678 baseline studies, 1,776 annotated nodules, and 60 cysts, accessed only via model-to-data.
2D Gum-Net predicts unsupervised dense elastic deformation fields
Zero-shot on real fingerprint benchmarks, structural correlation rises by up to 15.7% on FVC2004 after training only on synthetic impressions.
SurgKnowBench and SurgAgentGym test generalist models on surgery scenes
Agentic tool use mainly helps perception-heavy tasks; localization and expert-style assessment stay weak.
INFINITT PACS System Manager Contains Multiple Code Execution Vulnerabilities
INFINITT PACS System Manager versions 3.0.11.3 and 3.0.11.5 contain multiple code execution vulnerabilities that may allow local and remote attackers to execute arbitrary code.
Kev-Bio vision-language model for ultrasound biometry
Kev-Bio uses language model decisions to generate queries for landmark detection on the FU/GU Biometry 2026 dataset.
Pixel contrastive learning improves spine landmark detection on X-rays
A U-Net matches dense pixel embeddings to landmark prototypes, then uses learnable shape priors to resolve ambiguous vertebral corners.
HAD-VLM pairs diffusion with language for ultrasound measurements
It cuts four EchoNet-LVH measurement errors 21.4–37.6% versus a Vicuna baseline and lowers X-ray vertebral error from 11.4 to 6.7 pixels.
Untrained CNN and coefficient-space regularization for joint reconstruction in photoacoustic tomography
An untrained CNN with coefficient regularization jointly reconstructs photoacoustic pressure and sound speed.
Data-Efficient Classifier Pretraining Through Diffusion Models
DiffPT uses diffusion pretraining to improve low-data classifiers, especially on medical images and time series.
VascuSR: Iterative Anatomy-Aware Diffusion for CT Super-Resolution
VascuSR uses anatomy-aware diffusion and 3D vessel guidance to super-resolve CT scans with fewer false vessels.
VascuSR: Iterative Anatomy-Aware Diffusion for CT Super-Resolution
VascuSR super-resolves CT vessels with 3D anatomy guidance, cutting false vessel connections.
BigCell: Generating Gigapixel Whole-Slide Images
BigCell generates coherent gigapixel histopathology slides, and synthetic samples raise few-shot classification by up to 18%.
Efficient Secure Federated Learning via Information-Theoretically Secure Key Distribution: A Medical Imaging Case Study
Secure federated learning cuts information-theoretic key consumption about 35-fold for chest X-ray training.
From Scan to Treatment Plan, AI Helps Close Breast Cancer’s Deadliest Gaps
NVIDIA Inception startups use AI to speed breast cancer screening, risk assessment, and treatment planning.
2
Fiber-Resolved Microstructure Quantification from Multi-Shell Diffusion MRI using Detection Transformers
A Detection Transformer estimates multi-fiber diffusion MRI microstructure from standard multi-shell scans.
SentZero: An Enhanced Sentence-Centric Vision-Language Pretraining for Multi-Task Zero-Shot Chest X-Ray Analysis
SentZero improves zero-shot multi-task chest X-ray analysis with sentence-centric vision-language pretraining.
[SYSS-2026-068]: GDCM (Grassroots DICOM) - Stack-based Buffer Overflow (CWE-121)
SySS disclosed an unpatched stack-based buffer overflow in GDCM 3.3.0, a Grassroots DICOM library.
VCRE-Fib: View-Conditioned Regional Evidence for Fine-Grained Ultrasound Grading of Schistosoma japonicum-Associated Liver Fibrosis
VCRE-Fib improves ultrasound grading of schistosomiasis liver fibrosis using view-conditioned regional evidence.
LightMIS: Ultra-Lightweight Medical Image Segmentation Without a Stage-Wise Decoder
LightMIS segments medical images with 0.131M parameters and near Mobile U-ViT Dice while cutting compute sharply.
FleXray: Universal Clinical X-ray Segmentation
MIT CSAIL releases FleXray, a model that segments 60 anatomical structures in clinical X-rays.
Alibaba open-sources AI model that can detect cancer and nearly 150 conditions
Alibaba's Damo Academy open-sourced Damo Radar, a vision-language model that reads CT scans to detect nearly 150 abdominal conditions including cancers.