SearcharxivSearch

arXiv subjects

Noel DSouza

Publications and source records attributed to Noel DSouza.

2 recordsLinked to original sources

Decipher-MR: A Vision-Language Foundation Model for 3D MRI Representations

Magnetic Resonance Imaging is a critical imaging modality in clinical diagnosis and research, yet its complexity and heterogeneity hinder scalable, generalizable machine learning. Although foundation models have revolutionized language and vision tasks, their application to MRI remains constrained by data scarcity and narrow anatomical focus. We present Decipher-MR, a 3D MRI-specific vision-language foundation model trained on 200,000 MRI series from over 22,000 studies spanning diverse anatomical regions, sequences, and pathologies. Decipher-MR integrates self-supervised vision learning with report-guided text supervision to build robust representations for broad applications. To enable efficient use, Decipher-MR supports a modular design that enables tuning of lightweight, task-specific decoders attached to a frozen pretrained encoder. Following this setting, we evaluate Decipher-MR across disease classification, demographic prediction, anatomical localization, and cross-modal retrieval, demonstrating consistent improvements over existing foundation models and task-specific approaches. These results position Decipher-MR as a versatile foundation for MRI-based AI in clinical and research settings.

cs.CV

Experimental demonstration of acoustic wave induced magnetization switching of dipole coupled magnetostrictive nanomagnets for ultralow power computing

We report nanomagnetic switching with Acoustic Waves (AW) launched from interdigitated electrodes that modulate the stress anisotropy of elliptical cobalt nanoscale magnetostrictive magnets (340 nm x 270 nm x 12 nm) delineated on 128 degree Y-cut lithium niobate. The dipole-coupled nanomagnet pairs are in a single-domain state and are initially magnetized along the major axis of the ellipse, with their magnetizations parallel to each other. The magnetizations of nanomagnets having lower shape anisotropy are reversed upon acoustic wave propagation. Thereafter, the magnetization of these nanomagnets remains in the reversed state and demonstrate non-volatility. This executes a 'NOT' operation. This proof of acoustic wave induced magnetic state reversal in dipole-coupled nanomagnets implementing a 'NOT' gate operation could potentially lead to the development of extremely energy-efficient nanomagnetic logic. Furthermore, fabrication complexity is reduced immensely due to the absence of individual contacts to the nanomagnets, leading to lower energy dissipation

cond-mat.mes-hall