NSTL国家科技图书文献中心

1. Enhanced Deep Image Prior for Unsupervised Hyperspectral Image Super-Resolution NSTL国家科技图书文献中心

Jiaxin Li | Ke Zheng ... - 《IEEE Transactions on Geoscience and Remote Sensing》 - 2025,63 - 1～18 - 共18页 - 被引量：2

摘要： image prior (DIP) allows us to achieve unsupervised | establishment and a deep* image generation (DIG) stage for* | low-resolution hyperspectral image* (LrHSI), high* | -resolution multispectral image* (HrMSI), and corresponding* | high-resolution hyperspectral image* (HrHSI), the*

关键词： Electronics packaging | Hyperspectral imaging | Degradation | Image reconstruction | Zero shot learning | Generators | Training | Noise | Tensors | Estimation

2. A retinex inspired deep image prior model for despeckling and deblurring of aerial and satellite images using proximal gradient method NSTL国家科技图书文献中心

Architha Shastry | A. A. Bini ... - 《International journal of remote sensing》 - 2025,46(3/4) - 1432～1466 - 共35页

摘要： Image Prior model (DIP) addresses these issues by | ground truth and extensive training data. The Deep | performing restoration tasks using a single image, relying | despeckling and deblurring models using distinct image | ), Entropy, Global Contrast Factor (GCF), Natural Image

关键词： Deblurring | despeckling | proximal gradient descent ascent | deep image prior | retinex | total variation

3. High-Order DIP-VBTV: An Image Restoration Model Combining a Deep Image Prior and a High-Order Total Variation on Vector Bundles NSTL国家科技图书文献中心

Batard, Thomas - 《Journal of mathematical imaging and vision》 - 2025,67(2) - 共27页

摘要： image prior (DIP), yielding a variational model for | manifold. Then, we insert the high-order TV into the deep | image restoration. The proposed model can be viewed as | variational models for image* restoration: high-order TV* | edges and fine structures of the original image

关键词： Image restoration | Differential geometry | High-order total variation | Deep image prior

4. Fundamental study on improving the quality of X-ray fluorescence computed tomography images by applying deep image prior to projection images as a pre-denoising method NSTL国家科技图书文献中心

Kusakari, Sota | Sato, Kazuki ... - 《International journal of computer assisted radiology and surgery.》 - 2025,20(4) - 665～676 - 共12页

摘要： the quality of XFCT images by applying a deep image | of biological functions. Improvements in image | prior (DIP), which is a type of convolutional neural | image quality and was superior to the DIP post | and pixel values. Finally, image* quality of the*

关键词： X-ray fluorescence computed tomography | Deep image prior | Denoising | Biofunctional imaging

5. Analysis of Deep Image Prior and Exploiting Self-Guidance for Image Reconstruction NSTL国家科技图书文献中心

Shijun Liang | Evan Bell ... - 《IEEE transactions on computational imaging》 - 2025,11 - 435～451 - 共17页

摘要：The ability of deep* image prior (DIP) to* | measurements has made it popular in inverse problems in image | suggests that incorporating a reference image* as network* | input can enhance DIP's performance in image | network input and reconstructed image. We demonstrate

关键词： Image reconstruction | Electronics packaging | Training | Magnetic resonance imaging | Noise reduction | Kernel | Overfitting | Convolutional neural networks | Inverse problems | Information filters

6. Unsupervised Deep Learning for DAS-VSP Denoising Using Attention-Based Deep Image Prior NSTL国家科技图书文献中心

Yang Cui | Umair Bin Waheed ... - 《IEEE Transactions on Geoscience and Remote Sensing》 - 2025,63 - 1～14 - 共14页

摘要： method, denoising with a deep* image prior (DIP)-based* | deep learning (DL) methods, the proposed approach | Distributed acoustic sensing (DAS) has emerged | as a widely used technology in various applications | , including borehole microseismic monitoring, active source

关键词： Noise | Noise reduction | Training | Kurtosis | Feature extraction | Background noise | Noise measurement | Training data | Optical fibers | Data processing

7. Blind Image Deblurring with Noise-Robust Kernel Estimation NSTL国家科技图书文献中心

Chanseok Lee | Jeongsol Kim ... - 《Computer Vision - ECCV 2024,Part XX》 - European Conference on Computer Vision - 2025, - 188～204 - 共17页

摘要：-robust kernel estimation function and deep image prior | problem involving the retrieval of a clear image* and* | blur kernel from a single blurry image. The challenge | strongly noisy blurry images given a clear image* and* | generation of a clear image to leverage its natural image

关键词： Blind deblurring | Deep image prior | Gaussian noise

8. Deep Diffusion Image Prior for Efficient OOD Adaptation in 3D Inverse Problems NSTL国家科技图书文献中心

Hyungjin Chung | Jong Chul Ye - 《Computer Vision - ECCV 2024,Part LXXV》 - European Conference on Computer Vision - 2025, - 432～455 - 共24页

摘要：. In this work, we propose deep diffusion image prior | deep image prior. Under this framework, we propose an | , adaptation of the prior* is necessary when there exists a* | tasks from the generative prior* trained only with* | Recent inverse problem solvers that leverage

关键词： Diffusion models | Inverse problems | OOD adaptation

9. Spatial-temporal source term estimation using deep neural network prior and its application to Chernobyl wildfires NSTL国家科技图书文献中心

Antonie Brozova | Vaclav Smidl ... - 《Journal of hazardous materials》 - 2025,488(May 5) - 137510.1～137510.12 - 共12页

摘要： based on deep* image prior utilizing the structure of a* | the whole 5D tensor, some prior* information is* | deep neural network to regularize the inversion is | prior covariance structure in the source term. The | The source term of atmospheric emissions of

关键词： Atmospheric inversion | Spatial-temporal source | Deep image prior | Deep neural networks | Chernobyl wildfires

10. Temporal compressive complex amplitude imaging based on double random phase encoding NSTL国家科技图书文献中心

Xu, Ning | Qi, Dalong ... - 《Optics and Lasers in Engineering》 - 2025,184(Jan. Pt.1) - 1.1～1.10 - 共10页

摘要： a plug-and-play-based deep* image prior algorithm* | spatiotemporal information from a superimposed 2D image* of a* | Snapshot temporal compressive imaging offers a | potent method for capturing high-dimensional | dynamic scene. However, despite its notable bandwidth

关键词： Snapshot compressive imaging | Temporal imaging | Complex-amplitude imaging | Compressive sensing | Image reconstruction | Deep image prior | ALGORITHMS

摘要： image prior (DIP) allows us to achieve unsupervised | establishment and a deep image generation (DIG) stage for | low-resolution hyperspectral image (LrHSI), high | -resolution multispectral image (HrMSI), and corresponding | high-resolution hyperspectral image (HrHSI), the

关键词： Electronics packaging | Hyperspectral imaging | Degradation | Image reconstruction | Zero shot learning | Generators | Training | Noise | Tensors | Estimation

摘要： Image Prior model (DIP) addresses these issues by | ground truth and extensive training data. The Deep | performing restoration tasks using a single image, relying | despeckling and deblurring models using distinct image | ), Entropy, Global Contrast Factor (GCF), Natural Image

关键词： Deblurring | despeckling | proximal gradient descent ascent | deep image prior | retinex | total variation

摘要： image prior (DIP), yielding a variational model for | manifold. Then, we insert the high-order TV into the deep | image restoration. The proposed model can be viewed as | variational models for image restoration: high-order TV | edges and fine structures of the original image

关键词： Image restoration | Differential geometry | High-order total variation | Deep image prior

摘要： the quality of XFCT images by applying a deep image | of biological functions. Improvements in image | prior (DIP), which is a type of convolutional neural | image quality and was superior to the DIP post | and pixel values. Finally, image quality of the

关键词： X-ray fluorescence computed tomography | Deep image prior | Denoising | Biofunctional imaging

摘要：The ability of deep image prior (DIP) to | measurements has made it popular in inverse problems in image | suggests that incorporating a reference image as network | input can enhance DIP's performance in image | network input and reconstructed image. We demonstrate

关键词： Image reconstruction | Electronics packaging | Training | Magnetic resonance imaging | Noise reduction | Kernel | Overfitting | Convolutional neural networks | Inverse problems | Information filters

摘要： method, denoising with a deep image prior (DIP)-based | deep learning (DL) methods, the proposed approach | Distributed acoustic sensing (DAS) has emerged | as a widely used technology in various applications | , including borehole microseismic monitoring, active source