Arrate Muñoz‐Barrutia

The paper quantifies PanIN burden and pancreatic composition across age to identify high-risk demographics associated with increased precancer burden.

FilAdh is a framework for analyzing focal adhesions and actin stress fibers, achieving high accuracy in segmentation, reconstruction, and classification while revealing measurement biases under degradations.

The paper discusses AI4Life's Open Calls and Public Challenges, highlighting the state of bioimage analysis, tool availability, and the need for better data and infrastructure.

CODAvision provides a user-friendly interface for rapid, customizable medical image segmentation using deep learning, enabling non-experts to train and evaluate models across diverse imaging modalities.

This study shows that glioblastoma cells use transient cytoskeletal anisotropy to encode short-term mechanical memory, revealing how cancer cells adapt to mechanical stress.

This paper introduces NAFNet GAN, a bioimage restoration framework that achieves high fidelity and low latency, suitable for real-time microscopy applications.

This study uses Light Sheet Fluorescence Microscopy to analyze 3D kidney morphology in ADPKD mouse models, identifying biomarkers of tissue deformation.

BiaPy provides an open-source library for deep learning in bioimage analysis, offering tools for 2D/3D image data with multi-GPU support and customizable workflows.

The paper uses qEEG features from sleep stages to predict CSF oxidative biomarkers in early Alzheimer's disease with high accuracy.

This paper introduces an automated workflow using the SAM model to segment cysts in histological images of ADPKD, showing increased diseased tissue proportion over time and outperforming existing tools.

The paper introduces GAT, a tool for automated analysis of enteric neurons in gut tissue images, improving accuracy and throughput over manual methods.

The paper profiles the immune microenvironment of PanINs in 3D at single-cell resolution to understand inflammation's role in cancer progression and finds immune heterogeneity and correlations with PanIN grade.

This paper uses a generative AI model to improve 3D tissue imaging by interpolating and restoring images, enhancing resolution and reducing artifacts across multiple modalities and species.

This paper introduces enhancements to deepImageJ, a Fiji/ImageJ plugin, by integrating the Java Deep Learning Library (JDLL) to support multiple deep learning frameworks, enabling complex bioimage analysis tasks.

This paper develops a deep learning classifier to distinguish normal from abnormal cell division in mitosis videos, achieving high accuracy in classification.

This paper introduces HaPi, a method to automatically determine the handedness of CryoEM maps, achieving 89% accuracy on simulated data and identifying 11 incorrectly handed structures in a database.

The BioImage Model Zoo provides a community standard, infrastructure, and open repository for sharing, testing, and deploying pre-trained deep learning microscopy models across multiple bioimage analysis tools.

The paper proposes a method to adjust p-values based on sample size to better detect differences in large datasets and reduce uncertainty in statistical decisions.

NAP-CNB is a web server tool that predicts MHC-I-restricted T cell epitopes in mice using recurrent neural networks, achieving an AUC of 0.95 for H-2 peptide ligands.