SIA OpenIR  > 光电信息技术研究室
面向人眼感知的增强图像评价与图像优化方法研究
Alternative TitleResearch on Enhanced Image Evaluation and Image Optimization for Human Vision Perception
范晓鹏1,2
Department光电信息技术研究室
Thesis Advisor朱枫
ClassificationTP391.41
Keyword增强图像评价 图像优化 人眼视觉系统 人眼临界可见偏差 灰度差人眼探测概率
Call NumberTP391.41/F25/2017
Pages118页
Degree Discipline模式识别与智能系统
Degree Name博士
2018-05-18
Degree Grantor中国科学院沈阳自动化研究所
Place of Conferral沈阳
Abstract本文以增强图像评价指标作为目标函数,通过图像优化算法,实现目标函数最优化,最大化大脑接收的关于原图像的信息量。本文在增强图像评价和图像优化两方面开展了相关的研究工作,考虑到人在观察图像时关注的图像信息侧重不同,分别在图像细节、图像特征以及图像场景三个方面考虑大脑接收到的信息。以图像中相邻像素灰度大小关系代表图像的细节信息,局部像素灰度大小关系代表图像的特征信息,全局像素灰度大小关系代表图像的场景信息。因此从相邻像素灰度关系、局部像素灰度关系和全局像素灰度关系三个方面分别开展了增强图像评价和图像优化相关研究工作。本文的主要研究内容如下:1.针对相邻像素灰度关系,提出一种基于人眼视觉灰度差探测概率函数的增强图像评价算法,以相邻像素灰度差人眼探测概率的平均值作为增强图像的评价指标。以该评价指标作为目标函数,在相邻像素灰度关系保序前提下,通过迭代过程实现图像优化算法。2.针对局部像素灰度关系,提出一种以增强图像的对比度增益以及噪声估计为主,同时考虑人眼视觉系统的亮度阈值特性、对比度掩盖特性以及视觉通道内噪声等的增强图像评价指标。以该评价指标作为目标函数,将图像灰度映射函数分段表示,以映射函数节点变量作为优化变量,通过非线性优化过程实现图像优化算法。3.针对全局像素灰度关系,提出了以图像中人眼可分辨像素灰度差数量和人眼可感知图像互信息两种图像评价指标,并以全局单调递增灰度映射函数为基本前提,分别实现两种评价指标对应的图像优化算法。两种优化算法都采用动态规划策略,首先通过自上而下方法给出了最优解的证明过程,再通过自下而上方法完成了优化算法的具体实现。
Other AbstractUsing the enhanced image evaluation index as an objective function, an image optimization algorithm can be proposed to achieve the optimization of the objective function and maximize the information received by the brain regarding the original image. This paper has carried out related research work in both enhanced image evaluation and image optimization. Considering that people may pay attention to different image information in image observation, define the brain received image information in three aspects: image detail description, image characteristic description and image hierarchical description. It is considered that the relationship between the gray scales of adjacent pixels in the image reflects the details of the image, and the relationship between the gray scales of the local pixels reflects the characteristics of the image, and the relationship between the gray scales of the global pixels reflects the hierarchical information of the image. Therefore, the related research work of enhanced image evaluation and image optimization was carried out from the three aspects of the gray relationship between adjacent pixels, the relationship between local pixels and the global pixel gray relationship. The main research content of this paper is as follows: 1. Based on the gray relation of adjacent pixels, proposes an enhanced image evaluation algorithm in view of the human eye vision gray-scale detection probability function, and uses the average value of the gray-difference of adjacent pixels as the evaluation index of enhanced images. Based on the evaluation index, a corresponding objective function is proposed, and an image optimization algorithm is implemented through an iterative process based on preservation of the gray relation of adjacent pixels. 2. Based on the gray relation of local pixels, an enhanced image evaluation index is proposed, which focuses on the contrast gain and noise estimation of the enhanced image, and considers the brightness threshold value, contrast masking characteristics, and noise in the visual channel of the human visual system. Based on this evaluation index, a corresponding objective function is proposed, and the image optimization problem is then transformed into a multivariable nonlinear function minimization problem by simplifying the image grayscale mapping function into multi-node variables. 3. Based on the gray relation of global pixels, two kinds of image evaluation indexes are proposed considered global monotone increasing gray-scale mapping function. The evaluation indexes are the number of human eye-resolved pixel gray-scale differences and human-perceived image mutual information. Then the image optimization algorithms corresponding to the two evaluation indexes are realized. Both optimization algorithms provide a proof process of the optimal solution through the top-down approach of dynamic programming. Then the bottom-up approach is used to complete the implementation of the optimization algorithm.
Language中文
Contribution Rank1
Document Type学位论文
Identifierhttp://ir.sia.cn/handle/173321/21829
Collection光电信息技术研究室
Affiliation1.中国科学院沈阳自动化研究所
2.中国科学院大学
Recommended Citation
GB/T 7714
范晓鹏. 面向人眼感知的增强图像评价与图像优化方法研究[D]. 沈阳. 中国科学院沈阳自动化研究所,2018.
Files in This Item:
File Name/Size DocType Version Access License
面向人眼感知的增强图像评价与图像优化方法(11605KB)学位论文 开放获取CC BY-NC-SAApplication Full Text
Related Services
Recommend this item
Bookmark
Usage statistics
Export to Endnote
Google Scholar
Similar articles in Google Scholar
[范晓鹏]'s Articles
Baidu academic
Similar articles in Baidu academic
[范晓鹏]'s Articles
Bing Scholar
Similar articles in Bing Scholar
[范晓鹏]'s Articles
Terms of Use
No data!
Social Bookmark/Share
All comments (0)
No comment.
 

Items in the repository are protected by copyright, with all rights reserved, unless otherwise indicated.