Filters








154 Hits in 4.1 sec

Analysis of deep learning architectures for turbulence mitigation in long-range imagery

David Vint, Gaetano Di Caterina, John J. Soraghan, Robert A. Lamb, David Humphreys, Judith Dijk
2020 Artificial Intelligence and Machine Learning in Defense Applications II  
For the task of turbulence mitigation, real life data is difficult to obtain, as a clean undistorted image is not always obtainable.  ...  The use of deep learning for image processing has now become commonplace, with neural networks being able to outperform traditional methods in many fields.  ...  The networks under review were originally designed for tasks such as image denoising, super resolution or image deblurring.  ... 
doi:10.1117/12.2573927 fatcat:anz7ilqiybdyvflomztzoovuam

DiNTS: Differentiable Neural Network Topology Search for 3D Medical Image Segmentation [article]

Yufan He, Dong Yang, Holger Roth, Can Zhao, Daguang Xu
2021 arXiv   pre-print
Recently, neural architecture search (NAS) has been applied to automatically search high-performance networks for medical image segmentation.  ...  Existing methods either require long searching time for large-scale 3D image datasets, or are limited to pre-defined topologies (such as U-shaped or single-path).  ...  VNet [26] improves 3D UNet with residual blocks. UNet++ [47] uses dense blocks [13] to redesign skip connections.  ... 
arXiv:2103.15954v1 fatcat:a77g2iy6s5gftmgmz5dqzfn26e

Limited View Tomographic Reconstruction using a Cascaded Residual Dense Spatial-Channel Attention Network with Projection Data Fidelity Layer

Bo Zhou, S. Kevin Zhoua, James S. Duncan, Chi Liu
2021 IEEE Transactions on Medical Imaging  
In this work, we propose a cascaded residual dense spatial-channel attention network consisting of residual dense spatial-channel attention networks and projection data fidelity layers.  ...  In addition, our experimental results on Deep Lesion datasets demonstrate that our method is able to generate high-quality reconstruction for 8 major lesion types.  ...  [36] recently also demonstrated that spatial-channel attention can boost the image super-resolution performance.  ... 
doi:10.1109/tmi.2021.3066318 pmid:33729929 fatcat:dudx7o6zyjgkrekabq2kfjoiqy

Accurate Lung Nodules Segmentation with Detailed Representation Transfer and Soft Mask Supervision [article]

Changwei Wang, Rongtao Xu, Shibiao Xu, Weiliang Meng, Jun Xiao, Xiaopeng Zhang
2022 arXiv   pre-print
Then, a novel Network with detailed representation transfer and Soft Mask supervision (DSNet) is proposed to process the input low-resolution images of lung nodules into high-quality segmentation results  ...  Our DSNet contains a special Detail Representation Transfer Module (DRTM) for reconstructing the detailed representation to alleviate the small size of lung nodules images, and an adversarial training  ...  In this work, we conduct self-supervised training on a super-resolution model for medical images and smartly transfer the detailed representation knowledge in the trained super-resolution model to the  ... 
arXiv:2007.14556v3 fatcat:w4y7eltr7bfllcj7tmxovqmxki

MA-Unet: An improved version of Unet based on multi-scale and attention mechanism for medical image segmentation [article]

Yutong Cai, Yong Wang
2020 arXiv   pre-print
Although convolutional neural networks (CNNs) are promoting the development of medical image semantic segmentation, the standard model still has some shortcomings.  ...  We choose the best models in the medical image segmentation task for comparison, namely Attention U-net, CE-Net, UNet++, Unet and Channel-UNet.  ...  The dense hole convolution block captures broader and deeper semantic features by injecting four cascaded branches with multi-scale hole convolution, and uses shortcut connections to prevent the problem  ... 
arXiv:2012.10952v1 fatcat:6mwkaoa55jevtk5p4k3vf5on24

NTIRE 2020 Challenge on Perceptual Extreme Super-Resolution: Methods and Results [article]

Kai Zhang, Shuhang Gu, Radu Timofte, Taizhang Shang, Qiuju Dai, Shengchen Zhu, Tong Yang, Yandong Guo, Younghyun Jo, Sejong Yang, Seon Joo Kim, Lin Zha (+51 others)
2020 arXiv   pre-print
They gauge the state-of-the-art in single image super-resolution.  ...  The challenge task was to super-resolve an input image with a magnification factor 16 based on a set of prior examples of low and corresponding high resolution images.  ...  Extreme Super-Resolution Members: Younghyun Jo 1 (yh.jo@yonsei.ac.kr), Sejong Yang 1 , Seon Joo Kim 1,2 Affiliation: 1 Yonsei University 2 Facebook Title: Cascade SR-GAN for Extreme Super-Resolution Members  ... 
arXiv:2005.01056v1 fatcat:6nwj5ilbgbgjnmd6oy435hjdhi

Cell-Net: Embryonic Cell Counting and Centroid Localization via Residual Incremental Atrous Pyramid and Progressive Upsampling Convolution

Reza Moradi Rad, Parvaneh Saeedi, Jason Au, Jon Havelock
2019 IEEE Access  
Progressive up-sampling convolution gradually reconstructs a high-resolution feature map by taking into account short-and long-range dependencies.  ...  INDEX TERMS Cell counting, human embryonic cells, IVF, medical image analysis, deep learning.  ...  [38] came up with an interesting idea (sub-pixel convolution) to recover resolution in a single-image super-resolution scenarios.  ... 
doi:10.1109/access.2019.2920933 fatcat:w7a6kk7yqnfgpo6zribi7fpuvi

MR image reconstruction using deep learning: evaluation of network structure and loss functions

Vahid Ghodrati, Jiaxin Shao, Mark Bydder, Ziwu Zhou, Wotao Yin, Kim-Lien Nguyen, Yingli Yang, Peng Hu
2019 Quantitative Imaging in Medicine and Surgery  
CNN image reconstruction using Resnet yields comparable image quality to Unet with 10X the number of parameters. This has implications for training with significantly lower data requirements.  ...  Based on our assessments, we find that Resnet and Unet achieve similar image quality but that former requires only 100,000 parameters compared to 1.3 million parameters for the latter.  ...  Deep learning-based approaches are well-developed in computer vision tasks such as image super-resolution (5) (6) (7) (8) , denoising and inpainting (9) (10) (11) (12) , while their application to medical  ... 
doi:10.21037/qims.2019.08.10 pmid:31667138 pmcid:PMC6785508 fatcat:iofvymidenakpgiujovgubkdse

Rethinking Coarse-to-Fine Approach in Single Image Deblurring [article]

Sung-Jin Cho, Seo-Won Ji, Jun-Pyo Hong, Seung-Won Jung, Sung-Jea Ko
2021 arXiv   pre-print
Second, the single decoder of the MIMO-UNet outputs multiple deblurred images with different scales to mimic multi-cascaded U-nets using a single U-shaped network.  ...  Source code is available for research purposes at https://github.com/chosj95/MIMO-UNet.  ...  The use of multi-scale images as an input for a single U-Net has also proven to be effective in other tasks such as depth map super-resolution [6] and object detection [21] .  ... 
arXiv:2108.05054v2 fatcat:pt5plw5q6rhvvhnkrzuikyxtjm

Woodland Labeling in Chenzhou, China, via Deep Learning Approach

Wei Wang, Yujing Yang, Ji Li, Yongle Hu, Yanhong Luo, Xin Wang
2020 International Journal of Computational Intelligence Systems  
Secondly, this paper proposed a dense fully convolutional network (DFCN) which combines dense network with FCN model and achieves good semantic segmentation effect.  ...  DFCN method is used to label the woodland in Gaofen-2 (GF-2) remote sensing images in Chenzhou.  ...  There is also a lightweight network for image segmentation and classification [27] . In addition, the semantic segmentation also has similarities with image super-resolution reconstruction [28] .  ... 
doi:10.2991/ijcis.d.200910.001 doaj:a5d94b09bfc44d20af9dc9215ba39bb8 fatcat:mxpg5wto65ff7jtjlhkgzxny4i

Improved U-Net Remote Sensing Classification Algorithm Based on Multi-Feature Fusion Perception

Chuan Yan, Xiangsuo Fan, Jinlong Fan, Nayi Wang
2022 Remote Sensing  
This framework adds the channel attention module (CAM-UNet) to the original U-Net framework and cascades the shallow features with the deep semantic features, replaces the classification layer in the original  ...  The selection and representation of remote sensing image classification features play crucial roles in image classification accuracy.  ...  for super-resolution remote sensing images.  ... 
doi:10.3390/rs14051118 fatcat:a3w3pjsrofb5pfj5bhyl3nglje

A Deep Journey into Super-resolution: A survey [article]

Saeed Anwar, Salman Khan, Nick Barnes
2020 arXiv   pre-print
single image super-resolution.  ...  We introduce a taxonomy for deep-learning based super-resolution networks that groups existing methods into nine categories including linear, residual, multi-branch, recursive, progressive, attention-based  ...  Densely Connected Networks Inspired by the success of the DenseNet [60] architecture for image classification, super-resolution algorithms based on densely connected CNN layers have been proposed to  ... 
arXiv:1904.07523v3 fatcat:ovihxjadfja55hrytvhggj5c6q

Accurate tissue interface segmentation via adversarial pre-segmentation of anterior segment OCT images

Jiahong Ouyang, Tejas Sudharshan Mathai, Kira Lathrop, John Galeotti
2019 Biomedical Optics Express  
Optical Coherence Tomography (OCT) is an imaging modality that has been widely adopted for visualizing corneal, retinal and limbal tissue structure with micron resolution.  ...  To the best of our knowledge, this is the first work to show the potential of incorporating a cGAN into larger deep learning frameworks for improved corneal and limbal OCT image segmentation.  ...  Acknowledgements We would like to thank the Center for Machine Learning in Health (CMLH) at Carnegie Mellon University for providing a fellowship to TSM.  ... 
doi:10.1364/boe.10.005291 pmid:31646047 pmcid:PMC6788614 fatcat:rdohuctrsbhkjh2hu2fasjfie4

An Improved Boundary-Aware Perceptual Loss for Building Extraction from VHR Images

Yan Zhang, Weihong Li, Weiguo Gong, Zixu Wang, Jingxi Sun
2020 Remote Sensing  
(VHR) remote sensing images.  ...  However, searching for better CNN architectures is time-consuming, and the robustness of a new CNN model cannot be guaranteed.  ...  Acknowledgments: We thank Inria for providing the Inria Aerial Image Labelling Dataset and the WHU Aerial Dataset on their website (https://project.inria.fr/aerialimagelabelling/, http://study.rsgis.whu.edu.cn  ... 
doi:10.3390/rs12071195 fatcat:k4t6rdutq5eqthotvyv3wdrx3e

Guided Depth Map Super-Resolution using Recumbent Y Network

Tao Li, Xiucheng Dong, Hongwei Lin
2020 IEEE Access  
INDEX TERMS Depth map super-resolution, convolutional neural network, UNet network, atrous spatial pyramid pooling, attention mechanism.  ...  Depth map super-resolution (SR) can be used to enhance the resolution and improve the quality of depth maps.  ...  In order to solve this issue, depth map super-resolution (SR) is developed to enhance the resolution of depth maps.  ... 
doi:10.1109/access.2020.3007667 fatcat:2ncqv3rbxncwhesslnzyeylyva
« Previous Showing results 1 — 15 out of 154 results