基于互信息的复合遗传算法在医学图像配准中的应用

分析当前的医学图像配准技术及算法,选择基于体素相似性的图像配准方法,以互信息作为配准相似性测度,将混合遗传算法、自适应遗传算法和育种算法相结合,提出复合遗传算法,以提高遗传算法的鲁棒性,加速收敛过程,准确地实现多模医学图像配准,再利用此算法对无噪声图像和有噪声图像进行配准实验.实验结果证明了该算法的可行性,是一种值得推广的医学图像配准算法.     

基于视频图像的快速拼接算法研究

提出一种基于特征列的全自动无缝图像拼接算法.该算法以上一帧图像中特征最明显的一列像素为基础模板与下一帧图像匹配,并提出了自动检测误匹配以及对误匹配进行处理的方法.该算法最大的特点是其鲁棒性.对工程中需要提高拼接速度的环境,可以达到实时或准实时性的要求.     

基于特征点匹配的拼接算法研究与实现

针对图像间因具有旋转及光线强度差异等现象而导致的拼接效果不佳及拼接速度慢的问题,提出一种基于特征点的配准算法.该算法首先利用相位相关法确定图像重叠区域,然后采用改进Harris角点检测算法检测角点,再根据相似测度NCC(Normalized Cross Correlation)方法提取出匹配特征点对,最后用渐进渐出的方法实现拼接图像的融合.实验证明,该算法比传统算法在角点数目上减少了四分之一,有效地去除了拼接产生的鬼影现象.能有效地提高图像拼接的速度和精度.     

基于监督学习的前列腺MR/TRUS图像分割和配准方法

前列腺核磁超声图像配准融合有助于实现前列腺肿瘤的靶向穿刺。传统的配准方法主要是针对手动分割的前列腺核磁（Magnetic resonance, MR）和经直肠超声（Trans-rectal ultrasound, TRUS）图像上对应的生理特征点作为参考点,进行刚体或非刚体配准。针对超声图像因成像质量低导致手动分割配准效率低下的问题,提出一种基于监督学习的前列腺MR/TRUS图像自动分割方法,与术前核磁图像进行非刚体配准。首先,针对图像分割任务训练前列腺超声图像的活动表观模型（Active appearance model, AAM）,并基于随机森林建立边界驱动的数学模型,实现超声图像自动分割。接着,提取术前分割的核磁图像与自动分割的超声图像建立轮廓的形状特征矢量,进行特征匹配与图像配准。实验结果表明,本文方法能准确实现前列腺超声图像自动分割与配准融合,9组配准结果的戴斯相似性系数（Dice similarity coefficient, DSC）均大于0.98,同时尿道口处特征点的平均定位精度达1.64 mm,相比传统方法具有更高的配准精度。      

基于光流方向信息熵统计的微表情捕捉

以光流法为依据,提出了一种基于光流方向信息熵(entropy of oriented optical flow,EOF)统计的方法捕捉微表情关键帧.首先,采用改进的Horn-Schunck光流法提取视频流中相邻两帧图像的微表情运动特征;其次,采用阈值分析法筛选出投影速度模值较大的光流向量;之后,采用图像信息熵统计光流变化角度,进而得到视频序列的方向信息熵向量,通过对熵向量的分析,实现微表情关键帧捕捉;最后,本实验采用芬兰奥卢大学的SMIC微表情数据库和中国科学院心理研究所傅小兰的CASME微表情数据库作为实验样本,通过与传统的帧差法比较,证明了本文提出的算法优于帧差法,能够较好地表现出微表情变化趋势,为微表情识别提供基础.      

基于多分辨率分析的多传感器遥感图像融合方法

给出了一种基于非抽样Contourlet变换的遥感图像融合算法.根据分解后不同层次的特点采用不同的融合规则,对高频分量采用区域能量的融合规则,使得融合图像的光谱质量和空间质量得到均衡提高、实验表明,该算法可以更好地表现高分辨率全色图像的细节特征,同时在保持多光谱图像的光谱信息方面具有较大优势.     

基于梯度压缩的YOLO v4算法车型识别

为进一步提高智能交通系统对车辆及不同车型识别的泛化性、鲁棒性与实时性。根据检测区域的特征有针对性地构建数据集,改变余弦退火衰减（CD）学习率的更新方式,提出一种基于梯度压缩（GC）的Adam优化算法（Adam?GC）来提高YOLO v4算法的训练速度、检测精度以及网络模型的泛化能力。为验证改进后YOLO v4算法的有效性,对实际路况的车流进行采集后,利用训练完成的网络模型对不同密度车流进行定量的车型检测实验验证。经实验验证,改进后方法的整体检测结果要优于改进前,YOLO v4和YOLO v4 GC CD训练得到的网络模型在阻塞流样本下检测得到的准确率分别为94.59%和96.46%;在同步流样本下检测得到的准确率分别为95.34%和97.20%;在自由流样本下检测得到的准确率分别为95.98%和97.88%。      

基于互信息的混合蚁群算法及其在旅行商问题上的应用

为了提高蚁群算法的求解性能,从医学图像配准算法的思想出发,提出了一种基于互信息相似度的混合蚁群算法.为了表示最优路径和待配准路径之间的互信息熵,在蚁群算法的概率算子中增加了一个新的相似度影响因子,从而可以增加原算法的全局搜索能力,同时可以加速算法在解空间的搜索速度.将该算法应用在旅行商问题上,根据旅行商问题的特定环境,对混合蚁群算法的算式进行了一定程度的化简,使得算法在解决此类问题时,相应的时间复杂度降低.通过实验与多种传统算法进行对比,结果表明该改进算法在求解性能和跳出局部最小解方面都有一定程度的提高.     

菱镁矿矿山井筒视频图像匹配方法的改进研究

传统方法在菱镁矿矿山井筒视频图像的匹配过程中存在匹配精度差、误差率高等不足,不能满足视频图像匹配实时性的要求。提出一种基于优化的尺度不变特征转换算法,在保证图像配准精度的前提下,减少井筒视频图像的尺度变化,构造特征描述子并进行特征配准,实现对菱镁矿矿山井筒视频图像的精确匹配。     

基于椭圆对称方向矩的可见光与红外图像配准算法

针对图像匹配制导中异源图像匹配难度大的问题,提出一种基于椭圆对称方向矩的可见光与红外图像配准算法.基于最稳定极值区域提取异源图像中具有尺度和仿射不变特性的椭圆区域,利用聚类分割方法从中自动选取具有异源不变性的同质区域特征,用椭圆对称方向矩描述区域特征边界各方向上的相似程度,通过互相关性指标进行特征匹配,获取匹配特征对,利用匹配矫正策略减少误匹配.实验结果表明:较传统算法,进一步提高了可见光与红外图像关联特征的匹配效率,正确率超过了95%,计算时间缩短了近一半.基本满足图像匹配制导对匹配算法实时性好、匹配正确率高、抗干扰能力强等要求.      

Mapping image data to stereotaxic spaces: applications to brain mapping

A methodology for spatial normalization of image data is presented. This methodology is based on a map between homologous features of an individual brain and the target brain, which is used to drive a three-dimensional elastic warping transformation. Functional or structural information present in the original, nonnormalized images is preserved during this transformation. In particular, information such as the volume or the total amount of a radioactive agent in any brain region can be calculated directly from the normalized images. Moreover, subtle morphological characteristics of an individual brain are captured by the properties of the spatial transformation applied to that brain. Intersubject or interpopulation comparisons are performed by comparing the corresponding transformations.     

The registration of multiple medical images acquired from a single subject: why, how, what next?

This paper reviews some of the recent techniques which have been used to register multiple images of the same patient. Image registration is a problem which has been receiving significant attention from the medical image processing community in recent years. A successful image registration can aid in patient diagnosis, treatment assessment, image guided interventions, surgery planning and surgery. At present the majority of work has focused on rigid body transformations of images. We shall discuss some of the approaches used and outline a key automatic method in detail. In order to allow image registration of parts of the body which do not remain rigid, either due to patient movement or a change in pathology, nonlinear deformation techniques are being developed. We shall talk of the history of these methods before explaining deformations using landmarks and a recent extension to allow the definition of rigid structures in such warps in more detail. Validation of these methods is of great importance and we shall discuss work which has already been carried out on this topic for rigid body registrations as well as ideas for the validation of deformation algorithms.     

Automated 3-D registration of MR and CT images of the head

This paper discusses the application of voxel similarity measures in the automated registration of clinically acquired MR and CT data of the head. We describe a novel single-start multi-resolution approach to the optimization of these measures, and the issues involved in applying this to data having a range of different fields of view and sampling resolution. We compare four proposed measures of voxel similarity using the same optimization scheme when presented with 10 pairs of images with a range of initial misregistrations. The registration estimates are compared with those provided by manual point-based registration and evaluated by visual inspection to give an assessment of the robustness and accuracy of the different measures. One full-volume CT image set is used to investigate the performance of each measure when used to align truncated images from different regions in the head. The soft tissue correlation and mutual information measures were found to provide the most robust measures of misregistration, providing results comparable to or better than those from manual point-based registration for all but the most truncated image volumes.     

Medical image collection indexing: shape-based retrieval using KD-trees

The capacity to retrieve images containing objects with shapes similar to a query shape is desirable in medical image databases. We propose a similarity measure and an indexing mechanism for non-rigid comparison of shape which adds this capability to image databases. The (dis-)similarity measure is based on the observations that: (1) the geometry of the same organ in different subjects is not related by a strictly rigid transformation; and (2) the orientation of the organ plays a key role in comparing shape. We propose a similarity measure that computes a non-rigid mapping between curves and uses this mapping to compare oriented shape. We also show how KD-trees can index curves so that retrieval with our similarity measure is efficient. Experiments with real-world data from a database of magnetic resonance images are provided.     

基于三维仿射变换的数字图像置乱算法

针对数字图像的特点,基于有限整数域上的二维置乱变换、仿射变换和整数提升变换,提出了适用于任意大小、任意长宽比图像的三维置乱加密算法.考虑了变换矩阵中部分参数取负整数或小数的可行性,明确给出了参数的具体设置方法.该算法引入实数作为参数,扩展了参数选择范围;置乱像素位置的同时改变像素值,改善了置乱效果,加大了置乱周期,提高了数字图像的安全性.      

Finding parametric representations of the cortical sulci using an active contour model

The cortical sulci are brain structures resembling thin convoluted ribbons embedded in three dimensions. The importance of the sulci lies primarily in their relation to the cytoarchitectonic and functional organization of the underlying cortex and in their utilization as features in non-rigid registration methods. This paper presents a methodology for extracting parametric representations of the cerebral sulcus from magnetic resonance images. The proposed methodology is based on deformable models utilizing characteristics of the cortical shape. Specifically, a parametric representation of a sulcus is determined by the motion of an active contour along the medial surface of the corresponding cortical fold. The active contour is initialized along the outer boundary of the brain and deforms toward the deep root of a sulcus under the influence of an external force field, restricting it to lie along the medial surface of the particular cortical fold. A parametric representation of the medial surface of the sulcus is obtained as the active contour traverses the sulcus. Based on the first fundamental form of this representation, the location and degree of an interruption of a sulcus can be readily quantified; based on its second fundamental form, shape properties of the sulcus can be determined. This methodology is tested on magnetic resonance images and it is applied to three medical imaging problems: quantitative morphological analysis of the central sulcus; mapping of functional activation along the primary motor cortex and non-rigid registration of brain images.     

Effect of geometrical distortion correction in MR on image registration accuracy

In this article we investigate the effect of geometrical distortion correction in MR images on the accuracy of the registration of X-ray CT and MR head images for both a fiducial marker (extrinsic point) method and a surface-matching technique. We use CT and T2-weighted MR image volumes acquired from seven patients who underwent craniotomies in a stereotactic neurosurgical clinical trial. Each patient had four external markers attached to transcutaneous posts screwed into the outer table of the skull. The MR images are corrected for static field inhomogeneity by using an image rectification technique and corrected for scale distortion (gradient magnitude uncertainty) by using an attached stereotactic frame as an object of known shape and size. We define target registration error (TRE) as the distance between corresponding marker positions after registration and transformation. The accuracy of the fiducial marker method is determined by using each combination of three markers to estimate the transformation and the remaining marker to calculate registration error. Surface-based registration is accomplished by fitting MR contours corresponding to the CSF-dura interface to CT contours derived from the inner surface of the skull. The mean point-based TRE using three noncollinear fiducials improved 34%-from 1.15 to 0.76 mm-after correcting for both static field inhomogeneity and scale distortion. The mean surface-based TRE improved 46%-from 2.20 to 1.19 mm. Correction of geometrical distortion in MR images can significantly improve the accuracy of point-based and surface-based registration of CT and MR head images. Distortion correction can be important in clinical situations such as stereotactic and functional neurosurgery where 1 to 2 mm accuracy is required.     

Research progress of sinter tail sectional image based on computer vision

The section of the tail of sintering machine can directly reflect the rich information of sintering process control and sinter quality. How to use computer vision technology to extract the characteristics of the tail section and realize the accurate prediction and control of sinter quality is one of the key research contents of intelligent sintering. The image preprocessing methods such as denoising, segmentation and feature extraction of sinter tail section images were compared and analyzed comprehensively. Then the application situation of the image analysis technique for sintering machine tail section in the field of sinter quality prediction and control was analyzed from the aspects of FeO content in sintering ore, sintering end point, drum strength, distribution uniformity, sinter mixture moisture, etc. In addition, taking the prediction for FeO content in sintering ore as an example, the development and evolution rules of the study on sinter tail section image based on computer vision were revealed. The disadvantages of the current research and the development trend in the future were also pointed out.     

基于计算机视觉的烧结机尾断面图像研究进展

摘要：烧结机尾的断面可以最直接地反映烧结过程控制与烧结矿质量的丰富信息。如何利用计算机视觉技术对机尾断面的特征进行提取,并实现烧结矿质量的精准预报和控制,是智慧烧结的重点研究内容之一。首先对烧结机尾断面图像的去噪、分割和特征提取等图像预处理方式进行了全面的对比和分析,然后从烧结矿FeO含量、烧结终点、转鼓强度、布料均匀性和烧结混合料水分等方面剖析了烧结机尾断面图像分析技术在烧结矿质量预报和控制方面的应用情况。另外,以烧结矿FeO含量预报为例,揭示了基于计算机视觉的烧结机尾断面图像研究的发展和演变规律,并指明了当前研究的不足以及未来的发展趋势。