A new UNet variant reduces spectral artifacts in image transformations.
problem Spectral artifacts caused by traditional UNet upsampling layers.
method Introduced a Guided UNet (GUNet) architecture using a novel upsampling module.
result GUNet produces higher fidelity outputs in image transformations.
Trans-Unet predicts brain folding patterns from 3D point-clouds using novel 3D-to-2D transformation.
problem Challenges in learning high-fidelity 3D point-cloud features, including permutation invariance and fine-grained surface reconstruction.
method Transform 3D point-clouds into a 2D grid domain, then use a U-shaped hybrid model with CNNs and self-attention mechanisms.
result Trans-Unet achieves high-resolution predictions of brain patch growth, surpassing existing methods in fidelity and accuracy.
In this paper, we present UNet++, a new, more powerful architecture for medical image segmentation. Our architecture is essentially a deeply-supervised encoder-decoder network where the encoder and decoder sub-networks are connected through a series of nested, dense skip pathways. The re-designed skip pathways aim at r…
Deep neural network predicts traffic flow on city maps.
problem Short-term traffic flow prediction on high-resolution city maps.
method UNet-based deep convolutional neural network with densely connected layers.
result Best performance on the Traffic4cast challenge 2019.
Deep learning models, especially CNNs, can predict radio frequency power faster than traditional methods.
problem Accurate radio frequency power prediction for optimal transmitter location.
method Empirical analysis of deep learning models including CNNs and UNET variations for power prediction.
result Deep learning models, particularly CNNs, are effective and generalize well to new regions for power prediction.
A novel UNet detector detects sheep in UAV imagery.
problem Detecting small objects (sheep) in UAV imagery.
method Developed a novel dataset, used various object detectors, and evaluated their performance.
result UNet detector with weighted Hausdorff distance is best for sheep detection.
Global Planar Convolution boosts brain tumor segmentation by enhancing context perception.
problem Improving context perception in brain tumor segmentation networks.
method Introduced Global Planar Convolution module to enhance context aggregation in brain tumor segmentation.
result Global Planar Convolution eliminates the need for multiple representation levels in segmentation networks.
DL improves precipitation nowcasting from radar images.
problem Precise short-term precipitation predictions for extreme weather adaptation.
method Image-to-image translation using UNET CNN, compared to optical flow, persistence, and HRRR.
result UNET-based DL outperforms traditional models in 1 km x 1 km, 1 hour precipitation nowcasting.
A hybrid model combines diffusion and neural operator methods for stress prediction in hyperelastic materials.
problem Challenges in predicting stress fields in hyperelastic materials with complex microstructures.
method A hybrid surrogate framework combining a conditional denoising diffusion probabilistic model (cDDPM) and a modified DeepONet.
result The hybrid model consistently outperforms traditional methods by one to two orders of magnitude.
A deep learning algorithm for ECG segmentation.
problem ECG signal segmentation for various sampling rates and monitors.
method UNet-like neural network for adaptive and generalized ECG segmentation.
result F1-measures for ECG segmentation are at least 97.8%, 99.5%, and 99.9%.
ML-EM method speeds up diffusion model sampling.
problem Efficiently sampling from complex diffusion models.
method Multilevel Euler-Maruyama method with UNet approximations.
result Polynomial speedup in sampling from diffusion models.
ST-UNet models spatio-temporal graphs by pooling and unpooling operations.
problem Lack of effective means to extract dynamic features from spatio-temporal graphs.
method Designing a multi-scale architecture, Spatio-Temporal U-Net (ST-UNet), with paired sampling operations.
result Achieves substantial improvements in spatio-temporal prediction tasks.
Geometry-aware models improve cross-subject EEG decoding accuracy.
problem Strong inter-subject variability in motor imagery decoding.
method Discriminative Congruence Transform (DCT), Deep Linear DCT (DLDCT), Deep DCT-UNet (DDCT-UNet).
result Improves transductive cross-subject accuracy by 2-3%.
Time-aware deep learning methods improve spatial downscaling of atmospheric pollutants.
problem Transform coarse satellite data of atmospheric pollutants into high-resolution fields.
method Super-resolution deep residual networks and UNet architectures are extended with a temporal module encoding observation time.
result Temporal modules significantly improve downscaling performance and convergence speed.
Deep feature fusion improves mitosis counting accuracy.
problem Manual mitosis counting by pathologists is time-consuming and inconsistent.
method Combines Faster R-CNN for object detection with UNet segmentation features and RGB image features.
result Achieved an F-score of 0.508 on mitosis counting challenge dataset, outperforming state-of-the-art methods.
New method fills MS lesions more realistically and naturally.
problem Lesions in MS brain scans bias morphometric analyses.
method Non-local partial convolutions (NLPC) integrating Unet-like network and non-local module.
result NLPC generates inpainted regions that appear more realistic and natural.
Deep learning improves automatic image segmentation.
problem Automatic object localization and boundary delineation in images and medical scans.
method Proposed and evaluated novel dilated dense encoder-decoder architectures for salient object segmentation and lesion localization in medical images.
result Proposed architectures outperform state-of-the-art models in accuracy and efficiency.
Theory explains creativity in diffusion models generating novel images.
problem Diffusion models generate highly original images far from training data.
method Identified locality and equivariance as inductive biases to prevent optimal score-matching.
result Analytic models predict diffusion model outputs with high accuracy.
New CNN architecture improves pediatric image segmentation by homogenizing pose and size.
problem Challenges in segmenting pediatric images due to pose and size heterogeneity.
method Spatial Transformer Network (STN) for pose and scale invariance, combined with UNet for segmentation.
result Improved pediatric segmentation, especially renal tumor delineation, with accelerated processing.
Deep learning predicts lightning strikes with high accuracy.
problem Predicting thunderstorms and lightning strikes accurately.
method Convolutional neural network architecture inspired by UNet++ and ResNet.
result Probability of detection of more than 94% for lightning strikes within 15 minutes.
DeepMRSeg uses deep learning for brain tissue segmentation.
problem Automated segmentation of brain tissues and abnormalities.
method Modified UNet architecture with multi-scale feature extraction.
result DeepMRSeg achieves high accuracy on various brain segmentation tasks.
CAggNet improves medical image segmentation by fusing coarse and fine features.
problem Medical image segmentation accuracy and efficiency.
method Crossing Aggregation Network with nested skip connections and weighted aggregation.
result CAggNet achieves more accurate and efficient segmentation compared to existing methods.
A new model designs molecules with desired properties.
problem Finding molecules with optimal chemical or biological properties.
method A latent prompt Transformer model with three components: latent vector, molecule generation, and property prediction.
result The model achieves state-of-the-art performance on molecule design tasks.
Analyzes how diffusion models learn, revealing a spectral bias in structure mastery.
problem Understanding the learning dynamics and bias in diffusion models.
method Developed an analytical framework using a Gaussian-equivalence principle to solve gradient-flow dynamics and integrate probability-flow ODEs.
result Exposes a universal inverse-variance spectral law: high-variance structure is mastered faster than low-variance detail.
A new diffusion model generates structured tensors for high-dimensional data.
problem Generating a structured tensor with a target distribution.
method Tucker diffusion model with Tucker-Unet architecture.
result Generated tensors converge to the true data distribution at a rate dependent on tensor mode dimensions.
Deep-neural-network-based image reconstruction has demonstrated promising performance in medical imaging for under-sampled and low-dose scenarios. However, it requires large amount of memory and extensive time for the training. It is especially challenging to train the reconstruction networks for three-dimensional comp…
Diffusion models' consistency across splits explained by random matrix theory.
problem Consistency of diffusion models trained on non-overlapping subsets.
method Random matrix theory framework to quantify dataset effects on denoiser and sampling map.
result The theory explains and predicts cross-split disagreement in diffusion models.
SDIFT generates full-field dynamics from sparse, irregular data.
problem Modeling and reconstructing physical dynamics from sparse, off-grid observations.
method SDIFT uses a functional Tucker model and sequential diffusion for generating full-field evolution from irregular sparse observations.
result Significant improvements in reconstruction accuracy and computational efficiency compared to state-of-the-art approaches.
Deep learning ranks response surfaces for optimal stopping problems in finance.
problem Ranking response surfaces in stochastic control problems.
method Reformulate as image segmentation problem and apply deep learning algorithms.
result Deep learning provides an efficient method for solving optimal stopping problems.
Improved MRI head anatomy segmentation using deep learning with multiple priors.
problem Challenges in segmenting head anatomy in MRI, especially with lesions.
method Added three types of prior information to a 3D convolutional network: spatial priors, morphological priors, and spatial context.
result Multiprior network improves segmentation performance, especially for abnormal anatomies.
Synthetic learning improves neonatal brain MRI segmentation robustness.
problem Challenges in neonatal brain MRI segmentation due to image contrast and anatomical variations.
method Synthetic learning model trained on few T2-weighted volumes, then enhanced with motion artifacts and over-segmentation.
result Synthetic learning robust to image contrast and improves segmentation of both T1- and T2-weighted images.