98%
921
2 minutes
20
Computer-generated holography (CGH) offers a promising method to create true-to-life reconstructions of objects. While recent advances in deep learning-based CGH algorithms have significantly improved the tradeoff between algorithm runtime and image quality, most existing models are restricted to a fixed propagation distance, limiting their adaptability in practical applications. Here, we present a deep learning-based algorithmic CGH solver that achieves propagation-adaptive CGH synthesis using a spatial and Fourier neural operator (SFO-solver). Grounded in two physical insights of optical diffraction, specifically its global information flow and the circular symmetry, SFO-solver encodes both target intensity and propagation distance as network inputs with enhanced physical interpretability. The method enables high-speed 4 K CGH synthesis at 0.16 seconds per frame, delivering an average PSNR of 39.25 dB across a 30 mm depth range. We experimentally demonstrate various-depth 2D holographic projection and an adjustable multi-plane 3D display without requiring hardware modifications. SFO-solver showcases significant improvements in the flexibility of deep learning-based CGH synthesis and provides a scalable foundation to fulfill broader user-oriented requirements such as dynamic refocusing and interactive holographic display.
Download full-text PDF |
Source |
---|---|
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC12368012 | PMC |
http://dx.doi.org/10.1038/s41467-025-62997-z | DOI Listing |
Sci Rep
September 2025
Fukushima Renewable Energy Institute, Koriyama, Japan.
Ultra-fast charging stations (UFCS) present a significant challenge due to their high power demand and reliance on grid electricity. This paper proposes an optimization framework that integrates deep learning-based solar forecasting with a Genetic Algorithm (GA) for optimal sizing of photovoltaic (PV) and battery energy storage systems (BESS). A Gated Recurrent Unit (GRU) model is employed to forecast PV output, while the GA maximizes the Net Present Value (NPV) by selecting optimal PV and BESS sizes tailored to weekday and weekend demand profiles.
View Article and Find Full Text PDFIEEE Trans Biomed Eng
September 2025
Objective: Diffusion magnetic resonance imaging (dMRI) often suffers from low spatial and angular resolution due to inherent limitations in imaging hardware and system noise, adversely affecting the accurate estimation of microstructural parameters with fine anatomical details. Deep learning-based super-resolution techniques have shown promise in enhancing dMRI resolution without increasing acquisition time. However, most existing methods are confined to either spatial or angular super-resolution, disrupting the information exchange between the two domains and limiting their effectiveness in capturing detailed microstructural features.
View Article and Find Full Text PDFIEEE Trans Med Imaging
September 2025
Mammography is a primary method for early screening, and developing deep learning-based computer-aided systems is of great significance. However, current deep learning models typically treat each image as an independent entity for diagnosis, rather than integrating images from multiple views to diagnose the patient. These methods do not fully consider and address the complex interactions between different views, resulting in poor diagnostic performance and interpretability.
View Article and Find Full Text PDFJMIR Form Res
September 2025
Department of Critical Care Medicine, Guangdong Provincial People's Hospital (Guangdong Academy of Medical Sciences), Southern Medical University, Guangdong Provincial Geriatrics Institute, No. 106, Zhongshaner Rd, Guangzhou, 510080, China, 86 15920151904.
Background: Point-of-care ultrasonography has become a valuable tool for assessing diaphragmatic function in critically ill patients receiving invasive mechanical ventilation. However, conventional diaphragm ultrasound assessment remains highly operator-dependent and subjective. Previous research introduced automatic measurement of diaphragmatic excursion and velocity using 2D speckle-tracking technology.
View Article and Find Full Text PDFInt J Hyperthermia
December 2025
Department of Radiation Oncology Physics, University of Maryland, Baltimore, MD, USA.
Objective: To develop a deep learning method for fast and accurate prediction of Specific Absorption Rate (SAR) distributions in the human head to support real-time hyperthermia treatment planning (HTP) of brain cancer patients.
Approach: We propose an encoder-decoder neural network with cross-attention blocks to predict SAR maps from brain electrical properties, tumor 3D isocenter coordinates and microwave antenna phase settings. A dataset of 201 simulations was generated using finite-element modeling by varying tissue properties, tumor positions, and antenna phases within a human head model equipped with a three-ring phased-array applicator.