基于几何结构失真模型的图像质量评价研究

客观图像质量评价研究的目的是设计一种和视觉感知保持一致,且适用于各种失真模型的质量评价方法. 传统的结构相似度量质量评价方法忽视了自然图像本身的特点,不能很好地评判某些失真类型图像. 本文根据人眼视觉系统(Human visual system, HVS)在感知图像质量过程中的特点,探索自然图像的本征几何结构特征, 考虑像素点的方向失真、幅度失真和方差失真,提出了一种新型的基于图像几何结构失真模型的完全参考质量评价方法. 在标准数据库上的实验结果表明,本文方法适用于所有失真模型图像数据的质量评价, 计算复杂度相对较低,得到的图像客观评价结果和主观评价方法具有更好的一致性, 能够很好地反映人眼对图像质量的主观感受.     

基于单目视觉的水下机器人悬停定位技术与实现

以7000m载人潜水器的水下悬停为应用背景，以单目CCD摄像机为传感器，提出了基于模型的单目视觉定位和基于特征的视觉伺服两种水下机器人视觉悬停技术方案，分别适用于已知模型的观察目标和未知模型的观察目标两种情况．以自行研制的水下机器人控制系统实验平台为实验载体，利用前视摄像机和水下人工目标，建立了水下演示实验验证系统，并在室内实验水池中，分别实现了两种悬停方法的水下演示实验．演示实验表明：在两种视觉悬停方法的闭环控制下，水下机器人能够抵抗恒定水流干扰和人工位置扰动，很好地实现水下悬停作业．     

基于序列图像的视觉定位与运动估计

研究了移动机器人的视觉定位和目标的运动估计。采用单目视觉系统,借助人工标识物,由小孔成像模型及空间几何关系,推导出视觉测距模型,并实现了移动机器人的自定位和目标的定位。通过序列图像,应用基于特征的运动分析方法估计球体的运动参数,推导出移动机器人对运动目标的跟踪模型。球体定位实验结果表明:该方法的定位精度较高。     

基于HVS时频特性的视频质量评价方法

对于内容变化快的视频,现有仿真人类视觉系统的质量评价方法效果较差。因为主观评价者具有带通和掩蔽两种重要时域特征,现有方法中并未考虑这类与运动变化相关的视觉特性,导致客观与主观评价存在较大偏差。为提高运动变化快视频的评价准确性,利用统计学习的方法确定人类视觉阈值并建立模拟视觉系统的带通滤波模型;通过衰减权函数仿真人眼掩蔽特性。与基于信号特征的峰值信噪比(PSNR)方法、常权特征组合法、基于规则的变权特征评价法进行实验比较,在丢包率小于5%情况下获得最优评价效果。在带通模型的滤波作用下评价算法的执行效率得到提高。总之,该方法用简单方式有效地模拟复杂人类视觉特性,不仅提高评价性能而且降低计算复杂度。     

视觉检测中的光照强度优化研究

为了实现自适应控制光照强度,保障视觉检测系统获得高质量原始图像,研究了彩色图像成像感知模型,建立了图像灰度值与光源光照强度的关系,提出基于HIS彩色空间,单独使用I分量进行光照强度评价,给出了基于数字图像的清晰度和高频信息的光照强度评价函数,并设计了光照强度调控实验系统,通过实验分析了不同测度函数对光照强度变化的评价结果,取得了与人类主观评价一致的结果。实验结果表明,通过提出的方法控制光源强度,可以使视觉系统获得高质量的原始图像,保证检测的质量和精度。     

基于Navarro示意眼模型的视觉真实感绘制*

结合人眼光学建模和计算机图形学的真实感绘制技术,提出一种基于Navarro示意眼模型的人眼视觉真实感绘制方法.利用Navarro模型与传统的单透镜模型相比能够更精确地模拟人眼的特性,将Navarro模型引入视觉真实感绘制中;采用光线追踪方法,加入非球面折射面的计算,精确地模拟人眼的成像特性.实验结果表明此种方法能够更精...     

基于视觉注意力变化的网络丢包视频质量评估

针对网络中受丢包损伤的视频提出了一种基于视觉注意力变化的全参考客观质量评估方法.该方法基于视觉显著性检测在视频数据上的应用,考察受网络丢包失真影响的视频数据与标准参考数据在空间和时间上引起的视觉注意力变化,并根据此变化相应的视觉显著性在空间和时间上的差异,提出了一组客观质量评估方法.文中采用17个受丢包损伤的视频数据进行测试,并实施了主观评价实验作为评价标准.与传统的没有考虑人眼视觉显著特性的质量评估方法,以及目前主流的基于视觉显著区域/感兴趣区域对失真像素进行加权的方法进行对比,实验结果表明, 基于视觉注意力变化的方法较后两者与主观质量评估结果有更好的相关性, 能够更有效地评估丢包损伤视频的质量.     

基于人工路标和立体视觉的移动机器人自定位

针对室内移动机器人的自定位问题,提出一种基于人工路标和双目视觉的室内移动机器人自定位方法。首先设计了一种可扩展的彩色人工路标,并给出路标的编码方法;然后利用色彩空间变换,直线交比不变性以及自适应窗口实现路标检测与识别;最后在分析双目立体视觉模型的基础上建立起基于路标的双目立体视觉定位模型,实现移动机器人的准确定位。实验结果表明,路标对光照和视觉传感器的采集位置具有较强的鲁棒性,定位精度能够满足室内移动机器人的定位要求。     

一种图像背景视觉语义模板的生成方法

1.引言图像低层的物理视觉特征与人的高层认识之间不存在明显的直接联系,这就是视觉信息处理中的“语义鸿沟,这使得基于图像全局特征的检索结果与人的主观感觉大相径庭。要缓解“语义鸿沟”问题,一个直接的方法是在低层的视觉特征和高层的主观语义之间建立多个中间处理过程,使得两者能够有个渐进的过渡。这种分而治之的策略,需要保证每一步处理结果都要更加有利于主观语义的辨认,同时这些处理     

数字融合图像质量的视觉信息保真度客观评价方法

在实时数字图像融合系统中需要对融合图像的质量进行客观评价。针对传统客观评价方法存在与人眼视觉系统(Human Visual System,HVS)主观评价结果不相符的缺陷,提出了一种基于视觉信息保真度(Visual Information Fidelity,VIF)的融合图像质量客观评价方法。在假设图像源符合GSM(Gaussian Scale Mixture,GSM)模型,并将图像融合处理作为图像信号失真通道,而且考虑人眼视觉噪声特性的前提下,建立了基于VIF的数字融合图像质量客观评价理论模型,提出了一种融合图像质量评价指标FVIF。实验结果表明:该方法能够对融合图像质量进行综合评价,与其它方法相比,FVIF客观评价方法的性能更好,其评价结果与主观评价结果基本一致。     

A visual discomfort recognition model based on the fusion of multisource information

An experiment was performed to explore the influence of image attributes (white level, black level, color saturation, and sharpness) on visual discomfort, with measurements of the subjective questionnaire, eye movement, and electrocardiogram. The objective and subjective results showed that these image attributes influenced visual discomfort significantly. A visual discomfort recognition model based on Dempster–Shafer evidence theory was proposed according to the experiment results. With multiple features and weighting coefficients, the model could recognize visual discomfort states well.     

3维立体动画制作的视觉效果评价

目的 视觉效果评价是3维立体动画制作过程中不可忽略的一环。评价过程主要依靠专业人员的行业经验,受人员知识水平、测试环境等因素的影响。针对该问题,提出了一个主客观结合的评价模型。方法 首先建立了一个面向前期制作的三维动画场景数据集用以训练和测试,针对视觉效果评价的两个重要指标:立体感和视觉舒适度,进行主观实验得到相应的分数;提取全局视觉舒适度特征和感兴趣区域立体感特征,使用支持向量回归(SVR)方法,经过训练和测试得到舒适度评价模型和立体感评价模型。结果 通过将性能验证实验得到验证场景的主观分数与评价模型给出的结果进行比对,结果表明,运用评价模型得到的预测分数与观众主观分数基本一致,该模型可以对影响视觉效果的视觉舒适度和立体感予以5级量化评分。结论 本文所提出的视觉舒适度和立体感评价方法,能建立影响视觉效果的特征与主观评分间的关系,用得到的模型预测分数给制作人员一个及时直观的调节依据标准。     

一种自然场景图像的目标材质视觉特征映射算法

针对自然场景图像目标材质视觉特征映射中,尚存在特征提取困难、图像无对应标签等问题,本文提出了一种自然场景图像的目标材质视觉特征映射算法.首先,从图像中获取能表征材质视觉重要特征的反射层图像;然后,对获取的反射层图像进行前景、背景分割,得到目标图像;最后,利用循环生成对抗网络对材质视觉特征进行无监督学习,获得对图像目标材质视觉特征空间的高阶表达,实现了目标材质视觉特征的映射.实验结果表明,所提算法能够有效地获取自然场景图像目标的材质视觉特征,并进行材质视觉特征映射;与同类算法相比,具有更好的主、客观效果.     

Evaluation of stereoscopic visual fatigue in experts

Several objective or subjective measures have been employed to evaluate stereoscopic visual fatigue. Objective measures are more preferred for their capability to quantify the degree of human visual fatigue without being affected by individual variation. However, the combination of objective indicators, their sensibility in reflecting subjective fatigue and their applicability to different groups has rarely been studied. In our previous study based on non‐expert experiment, we proposed a method combining subjective assessment and objective measures to evaluate stereoscopic visual fatigue and an objective visual fatigue model (non‐expert model). In this study, we use the same procedure to conduct a contrast experiment in 2D and 3D conditions, but only with expert viewers. The results in 2D condition show that the two objective indicators are usable for visual fatigue evaluation. The difference between the results in 2D and 3D conditions shows that the stereo 3D can cause much more fatigue and the indicators can efficiently identify the extra fatigue with satisfactory accuracy. An objective visual fatigue model (expert‐only model) is developed based on the data obtained in the 3D condition and is evaluated with comparison to the non‐expert model. Both visual fatigue models show high correlation with the subjective data. The method we proposed is reasonably insusceptible to different viewer groups, and the results of repeated experiments are consistent. It is a promising approach to evaluate visual fatigue efficiently and objectively.     

Artificial neural network visual model for image quality enhancement

An artificial neural network visual model is developed, which extracts multi-scale edge features from the decompressed image and uses these visual features as input to estimate and compensate for the coding distortions. This provides a generic postprocessing technique that can be applied to all the main coding methods. Experimental results involving postprocessing of the JPEG and quadtree coding systems show that the proposed artificial neural network visual model significantly enhances the quality of reconstructed images, both in terms of the objective peak signal-to-noise ratio and subjective visual assessment.     

Perceived Visual Complexity and Visual Search Performance of Automotive Instrument Cluster: A Quantitative Measurement Study

Advancements in technology have spurred the development of new in-vehicle applications. Drivers are faced with different driving contexts due to an increase in the number of devices that provide a wealth of diverse information. However, such a scenario can cause drivers to become distracted. Therefore, research on how the presentation of visual information can affect drivers’ performance is important. In this study, an analysis of quantifiable measurements that affect drivers’ perception of visual complexity and visual search performance was conducted. A questionnaire was administered to assess subjective perception of visual complexity, and a user experiment using eye tracking was designed to explore participants’ visual search performance. The results of subjective visual complexity perception and visual search performance suggested that some objective measurement variables were significantly related only to perceived visual complexity, whereas others affected both subjective and behavioral measurements. Thus it is possible to predict which quantifiable measurement variables affect subjective perception of visual complexity and which affect visual search performance. Therefore, this study allows understanding and explaining of perception of visual complexity by quantifiable measurements and the different ways by which these measurements affect visual search performance.     

A computational model of perceptual saliency for 3D objects in virtual environments

When giving directions to the location of an object, people typically use other attractive objects as reference, that is, reference objects. With the aim to select proper reference objects, useful for locating a target object within a virtual environment (VE), a computational model to identify perceptual saliency is presented. Based on the object’s features with the major stimulus for the human visual system, three basic features of a 3D object (i.e., color, size, and shape) are individually evaluated and then combined to get a degree of saliency for each 3D object in a virtual scenario. An experiment was conducted to evaluate the extent to which the proposed measure of saliency matches with the people’s subjective perception of saliency; the results showed a good performance of this computational model.     

Color management of reflective‐type LCDs in terms of adaptation of the human visual system to light‐source variation

We have investigated the color management, in terms of the color adoption property of the human visual system, of a reflective‐type TFT‐LCD (R‐LCD). Since the R‐LCD depends on ambient light as the light source, it is expected that the colorimetric color on the R‐LCD must be changed if the source of the ambient light is changed. However, due to the adaptation property of the human visual system, the eye does not perceive colorimetrically corrected colors as the same color even for an R‐LCD. In this research, first, we conducted a subjective experiment to obtain the RGB code value that is required in order to display a corresponding color on the R‐LCD under varying ambient‐light conditions. The result of the experiment shows that the corresponding color of the experimental results was reasonably approximated by the color obtained by using the von Kries model. Secondly, we proposed a color‐compensating mechanism that is described as a cascaded simple 3 × 3 linear matrix. Actual colors displayed are adjusted according to the ambient light. The evaluation of the picture quality of the R‐LCD showed that the proposed model is effective.     

A Visual Database Environment for Scientific Research

This paper describes a visual database environment designed to be used for scientific research in the imaging sciences. It provides hierarchical relational structures that allow the user to model data as entities possessing properties, parts and relationships, and it supports multi-level queries on these structures. A schema constructor interface allows users to define for each structure, not only its components, but also its visualization, which is built from its components using graphical primitives. Finally, an experiment management subsystem allows users to construct and run computational experiments that apply imaging operators to data from the database. The experiment management system keeps track of the experimental procedures developed by the user and the results generated by executing these procedures.     

基于人眼视觉特性的图像视觉质量评价方法

为了更好地评价图像质量,在研究人眼视觉模型和人眼视觉特性的基础上,提出一种自顶向下的无参考图像客观质量评价方法。对待评价图像进行分块,并在每个色彩通道中对子块进行计算,输出平均视觉特性值,使评价模型既符合客观质量评价的优点,又兼顾人眼视觉感受,符合人眼主、客观一致性。选取美国TEXAS大学LIVE的图像数据库中图像的主观评价值对方法进行了实验证明,实验结果表明所提方法与主观评价取得了较好的一致性。