相关实验视频

Updated: Jan 16, 2026

Photorealistic Learned Landscapes for Augmented Reality

Photorealistic Learned Landscapes for Augmented Reality

Published on: June 27, 2025

图像修复是由双尺度前置驱动的

Weimin Yuan¹, Cai Meng¹, Xiangzhi Bai¹

¹Image Processing Center, Beihang University, Beijing, 100191, China.

Neural networks : the official journal of the International Neural Network Society

|September 29, 2025

概括

本研究引入了一种新的图像修复 (IR) 双尺度先前 (DSP) 模型. 作为一种有效的IR方法,DSPIR通过结合非学习和基于学习的先验来增强图像剥离和绘制.

科学领域:

计算机视觉计算机视觉
图像处理图像处理
机器学习机器学习

背景情况:

先进的成像技术增加了对高质量的图像的需求.
图像因噪音和数据丢失而恶化,影响质量.
现有的图像修复方法在灵活性和通用性方面存在局限性.

研究的目的:

为图像恢复引入一种新的双尺度前置 (DSP) 模型.
整合非学习和基于学习的优先课程的优势.
使用DSP模型开发一种有效的图像恢复方法 (DSPIR).

主要方法:

开发了一种双级预测 (DSP) 模型,结合了组级物理预测 (NSS) 和图像级深度排泄预测.
将DSP纳入后期最大限度 (MAP) 原则,以创建DSPIR.
使用乘数的交替最小化和交替方向方法解决了DSPIR.

主要成果:

DSP模型有效地保留了边缘,并消除了噪音.
DSPIR在各种降解类型中表现出强度.
广泛的评估表明,DSPIR在无雾化和涂漆方面优于最先进的方法.

结论:

双尺度前模型为图像恢复提供了一个强大的方法.
DSPIR 在图像无色化和inpainting方面实现了卓越的性能.
拟议的方法显示了与现有技术相比的显著改进.

关键词:

双尺度的先前情况.图像退化图像退化图像恢复图像恢复

更多相关视频

Measuring the Shape and Size of Activated Sludge Particles Immobilized in Agar with an Open Source Software Pipeline

Measuring the Shape and Size of Activated Sludge Particles Immobilized in Agar with an Open Source Software Pipeline

Published on: January 30, 2019

Detection of Architectural Distortion in Prior Mammograms via Analysis of Oriented Patterns

Detection of Architectural Distortion in Prior Mammograms via Analysis of Oriented Patterns

Published on: August 30, 2013

相关概念视频

Downsampling

Downsampling

When considering a sampled sequence with zero values between sampling instants, one can replace it by taking every N-th value of the sequence. At these integer multiples of N, the original and sampled sequences coincide. This process, known as decimation, involves extracting every N-th sample from a sequence, thereby creating a more efficient sequence.
The Fourier transform of the decimated sequence reveals a combination of scaled and shifted versions of the original spectrum. This...