Tag: deep learning

A retrospective imaging study found that six AI-assisted CT measures of left atrial shape are mathematically distinct and vary in ...

A 2026 review finds that dataset diversity, annotation quality and external validation remain major barriers to reliable clinical generalization of ...

A retrospective study of 50 radiotherapy patients found that daily machine-performance data and synthetic collimator-error training improved predicted treatment-delivery quality ...

A weighted training method that distinguishes likely shortcut examples improved AI robustness across four image and language benchmarks without adding ...

A dual-modality artificial-intelligence model that analyzed both fundus photographs and optical coherence tomography (OCT) scans reported high performance when classifying ...

A four-channel neural network using co-acquired atomic force microscopy signals improved image restoration, outperformed larger single-channel models and reduced training-data ...

A deep-learning framework classified sleep stages from single-channel EEG with reported accuracies of 84.5% to 87.0% across three public datasets, ...

A benchmark of 13 models on more than 47,000 English tweets found that TF-IDF-based Extra Trees led overall, while class-level ...

UAC-Net combined contextual attention and multi-scale feature fusion in a shared neural network that classified three brain-tumor types and delineated ...

In a 100-person multiple sclerosis imaging cohort, an automated lesion-segmentation and diffusion-tensor pipeline found that reference-like white matter within lesions ...

A context-aware deep-learning model classified rodent sleep stages across unseen animals and retained near-baseline performance after training only on vehicle-control ...

A self-supervised, physics-driven deep-learning method produced reliable functional brain maps with 20-fold accelerated MRI acquisitions in retrospective and prospective evaluations.

A study in Delhi, Mysore and Surat found that clustering road segments before training deep-learning models improved bus arrival-time predictions ...

A new deep learning architecture called TexTNet integrates local binary pattern texture features into a Swin Transformer's attention mechanism to ...

Extended depth of field and 3D imaging have enabled the description of two new endangered grasshopper species in the cryptic ...

A new review in PhotoniX Synergy details how machine learning, generative models, and large language models are transforming the design ...

A new deep learning study projects that leishmaniasis incidence in Ethiopia will fall to about 84 cases per 100,000 by ...

A hybrid deep learning framework combining Siamese networks and Vision Transformers achieves up to 95 percent accuracy in verifying parent-child ...

New research from IIT Kharagpur shows that federated learning drives models toward flatter loss minima that suppress memorization, and introduces ...

A new AI framework called GECST combines generative adversarial networks with classical compositional rules to restore China's fading Dunhuang murals ...