Hybrid AI Model Retrieves Medical Images Even When Queries Are Blurred or Noisy
A hybrid CNN–transformer model produces 128-dimensional medical image embeddings that stay robust under noise, blur, cropping, and occlusion across four ...
Tag: Medical Imaging
Lead-Loaded Plastic Scintillator Shows Promise for Faster, Cheaper PET Detectors in Simulation
Geant4 Monte Carlo simulations show that a 5 wt% lead-loaded plastic scintillator produces a measurable 511 keV photopeak with detection ...
Zero-Watermarking Scheme Aims to Protect Color Medical Images Without Altering Them
A new zero-watermarking scheme for color medical images uses VGG19 feature extraction, DCT and cyclic chaos encryption to verify copyright ...
AI Takes On the Flawed Math of Measuring Leaky Heart Valves
A new review in Cardiovascular Ultrasound explains how artificial intelligence tools, including deep learning PISA segmentation and 3D Auto Color ...
Lightweight Dual-Branch AI Reads Brain Scans to Spot Alzheimer’s Stages
A lightweight dual-branch AI framework called NeuroCST classified Alzheimer's disease stages from MRI scans with 98.42 percent accuracy while using ...
MRI Deep Learning Model Distinguishes Liver Cancer Types Before Surgery
A multi-phase MRI deep learning model combined with clinical predictors differentiated hepatocellular carcinoma from mass-forming intrahepatic cholangiocarcinoma with AUCs of ...
Three-Branch AI Fusion Reaches 99.3% Accuracy in White Blood Cell Classification
A hybrid deep learning model fusing MobileNetV3, DenseNet121, and DINOv2 achieved up to 99.3% accuracy in classifying five types of ...
Deep learning maps ventilator-associated pneumonia on CT and predicts mortality risk
A retrospective study of 512 ICU patients reports a multimodal deep-learning framework that segments CT lesions in ventilator-associated pneumonia and ...
Federated learning framework aims to stabilize colorectal cancer CT classification across hospitals
A new personalized federated learning framework uses Transformer-based dynamic aggregation and adaptive regularization to improve colorectal cancer CT image classification ...
Convolution-Inspired Spiking Neural P Circuit Achieves Lossless 75 Percent Image Resizing in Reversible Data-Hiding Framework for DICOM Medical Images
A convolution-inspired Spiking Neural P circuit encodes halftone binary images into colored spikes with a 75 percent size reduction and ...
Explainable AI lags in stroke lesion deep learning, review finds
A systematic review finds hybrid CNN-transformer models lead stroke lesion analysis, but explainable AI appears in only 11 percent of ...
MRI-based radiomics nomogram predicts axillary lymph node status before breast cancer surgery
A retrospective study of 201 patients built a nomogram combining ultrafast DCE-MRI and ADC radiomics features with clinical risk factors ...
Intestinal ultrasound score tracks Crohn’s disease activity with high accuracy
A prospective study of 28 Crohn's disease patients found that a standardized intestinal ultrasound score correlated strongly with clinical activity ...
Decades of AI in Prostate Cancer Show Technical Gains but Clinical Utility Remains Elusive
A review of three decades of artificial intelligence in prostate cancer demonstrates that while technical performance has advanced significantly, definitive ...
Explainable AI Could Help Diagnose Neglected Tropical Diseases in Darker Skin, Review Finds
A structured narrative review finds that explainable AI techniques may improve transparency in diagnosing skin-manifesting neglected tropical diseases, while deep ...
AI Model Impression-R1 Matches Radiologists in Multicenter Impression Writing Test
A domain-specialized reasoning model called Impression-R1 matched or outperformed general-purpose AI and participating radiologists on radiological impression generation across 11,884 ...
Deep Learning on Liver MRI Shows Promise for Detecting Liver Cancer, but Evidence Remains Thin
A systematic review finds that deep learning on liver MRI reaches pooled sensitivity of 0.862 and specificity of 0.916 for ...
Neural network turns a thin scintillator into a sharp low-energy gamma-ray camera
A thin monolithic CsI(Tl) crystal read out by an 8 by 8 silicon photomultiplier array resolves 59.5 keV gamma-ray positions ...
Ultrasound tracks knee cartilage loss but falls short of MRI precision
A cross-sectional study of 60 knee osteoarthritis patients found ultrasound strongly correlated with MRI in measuring medial femoral cartilage thickness ...
MRI and Ultrasound Disagree on Crohn’s Disease Measurements, Major Study Finds
A large multicentre analysis of the METRIC trial found that MR enterography and intestinal ultrasound agree well on a global ...
POPULAR NEWS
-
Metagenomic Sequencing Identifies Pathogens in 40.6 Percent of Suspected CNS Infections
-
AI Models Show Promise in Detecting Early Eating Disorders
-
Pediatric cytomegalovirus retinitis may follow distinct trajectory after stem cell transplantation
-
German Midwifery Educators Navigate Evidence-Based Practice Shift Across University Programmes