The field of medical imaging is rapidly advancing with the integration of artificial intelligence (AI) and deep learning techniques. Recent developments have focused on improving the accuracy and interpretability of AI models in various medical imaging tasks, including disease diagnosis, image segmentation, and report generation. Notably, the use of multimodal approaches, incorporating both imaging and clinical data, has shown promise in enhancing diagnostic precision and reducing inter-observer variability. Furthermore, the development of lightweight and efficient models has enabled the deployment of AI-assisted diagnostic tools in resource-constrained settings. Key areas of innovation include the application of foundation models to radiology tasks, the development of explainable AI frameworks for medical image analysis, and the creation of large-scale datasets for training and evaluating AI models. Particular noteworthy papers include the introduction of Pillar-0, a radiology foundation model that has achieved state-of-the-art performance on several tasks, and the development of OncoVision, a multimodal AI pipeline for enhanced breast cancer diagnosis. Additionally, the proposal of MedROV, a real-time open-vocabulary detection model for medical imaging, has demonstrated significant improvements over previous state-of-the-art models.
Advances in Medical Imaging and AI-Assisted Diagnostics
Sources
Multimodal AI for Body Fat Estimation: Computer Vision and Anthropometry with DEXA Benchmarks
Explainable Deep Learning for Brain Tumor Classification: Comprehensive Benchmarking with Dual Interpretability and Lightweight Deployment
Pillar-0: A New Frontier for Radiology Foundation Models
Toward explainable AI approaches for breast imaging: adapting foundation models to diverse populations
A Lightweight, Interpretable Deep Learning System for Automated Detection of Cervical Adenocarcinoma In Situ (AIS)
Less Is More: An Explainable AI Framework for Lightweight Malaria Classification
Radiation-Preserving Selective Imaging for Pediatric Hip Dysplasia: A Cross-Modal Ultrasound-Xray Policy with Limited Labels
Health system learning achieves generalist neuroimaging models
CellFMCount: A Fluorescence Microscopy Dataset, Benchmark, and Methods for Cell Counting
OncoVision: Integrating Mammography and Clinical Data through Attention-Driven Multimodal AI for Enhanced Breast Cancer Diagnosis
CountXplain: Interpretable Cell Counting with Prototype-Based Density Map Estimation
Vision-Language Models for Automated 3D PET/CT Report Generation
MedROV: Towards Real-Time Open-Vocabulary Detection Across Diverse Medical Imaging Modalities
UruDendro4: A Benchmark Dataset for Automatic Tree-Ring Detection in Cross-Section Images of Pinus taeda L
BUSTR: Breast Ultrasound Text Reporting with a Descriptor-Aware Vision-Language Model