图像

The population consists of a lot of individuals.The image with better quality is obtained by the evolutions of populations.

图像恢复的目标是对退化的图像进行处理，使它恢复成没有退化的理想图像。

Methods to 15 Han face bone fracture sufferer line thin layer continuous after scan, make use of the work station carry on empress the picture a processing and choose to use variety software technique, many flat surface reconstruction, the 3D surface cover up manifestation(3 D-SSD) and capacity reappear method etc. reorganization observation, reconstruction Han face the 3D image of the bone fracture, acquire the best picture photograph, and with CT crossing the picture carry on more analysis.

对15例颌面部骨折患者行薄层连续扫描后，利用工作站进行图像后处理，选用多种软件技术，多平面重建、三维表面遮盖显示（3D-SSD）及容积再现法等重组观察，重建颌面部骨折的三维影像，获得最佳图像照片，并与CT横断图像进行比较分析。

By the collected database of images,use the Five-line Seeking-points method to analysis the size and determine the denomination;based on the characters of gray scale of images,take and analyze for judging the face and obverse;Moreover,the gridding metho.

通过对采集的图像数据，采用五线找点算法对纸币大小进行判定纸币的面额；根据加权灰度值提取纸币图像的特征区域判正反；采用网格划分法统计背景点与纸币图像点的数量关系判定纸币的残缺率。

First of all two-dimensional bar code of the positioning and segmentation algorithm using Hough Transform and Sobel edge detection of the bar code images collected from the original image to effectively separate and then analyzes the barcode image through the optical system of the noise model, put forward a Calculation of point spread function of standard deviation of the algorithm uses Flourier transform adaptively select the threshold to remove noise caused by invalid border, bar code and thus be the basic module.

首先探讨了二维条码的定位与分割算法，利用Hough变换与Sobel边缘检测把条码图像从原始采集的图像中有效地分割出来然后分析了条码图像经过光学系统的噪声模型，提出了一种计算点扩展函数标准方差的算法采用Flourier变换自适应地选取阈值去除噪声导致的无效边界，从而得到条码的基本模块。

We define some perceptive parameters, such as the change of texture character,histogram,frequency spectrum ,knowledge of image,and we calculate the perceptive information using sigmoid function.

我们从图像灰度特征，直方图，频带，图像理解的变化等方面初步定义了图像感知信息参量，并利用sigmoid函数给出了感知信息的初步计算方法，实验结果初步说明了感知信息的有效性和可行性

We define some perceptive parameters,such as the change of texture character,histogram,frequency spectrum,knowledge of image,and we calculate the perceptive information using sigmoid fu...

我们从图像灰度特征，直方图，频带，图像理解的变化等方面初步定义了图像感知信息参量，并利用sigmoid函数给出了感知信息的初步计算方法，实验结果初步说明了感知信息的有效性和可行性

First, it gests the bandwidth parameter and center frequency of Gabor filter based on human vision and the size of texture image, constructing the multichannel Gabor filters and extracting the Gabor multitemporal and multidirectional energy features of image in the time and frequency domain and implementing nonlinear transform and Gauss lowpass filtering in the feature space.

首先根据人类视觉特性选定Gabor滤波器的频率带宽参数，利用图像大小确定滤波器的中心频率，构造一组多通道的Gabor滤波器在时域和频域中提取图像多方向和多分辨率的Gabor能量特征，对特征空间进行非线性变换和Gauss滤波处理，最后通过Kmeans方法实现纹理图像的分割。

Based on the two face images of the same person at front and half-profile view and an MPEG4 3D generic face model, we estimate the poses of the face at these two images and establish their 2D mesh models. Multiple views of the same face can be synthesized by image interpolation and warping, meanwhile occlusion problems are solved reasonably.

基于同一人脸的正面和半侧面图像，引入三维通用人脸模型估计其初始朝向，分别建立两图像的二维网格模型；然后采用图像插值和变形技术合成不同朝向人脸，并合理地解决遮挡问题。

Finally,we compare the similarity of the color and texture histograms between the query example and each one in the image database.

随着遥感图像数量的急剧增加，如何进行高效检索已成为遥感图像信息提取和共享的瓶颈问题，基于内容的遥感图像检索因此逐渐成为了研究热点。

Finally, according to the needs of online identification, the thesis makes the solutions of access to moving image real-time by machine vision techenology. IC printed mark automatic identification system base on FV1100 image processing equipment is built.

最后，针对工业字符在线识别中获取运动图像的需求，根据机器视觉知识，设计了实时获取运动图像的解决方案，构建了基于FV1100图像处理设备的芯片标识识别实验系统。

The split between the two groups can hardly be papered over.

这两个团体间的分歧难以掩饰。

This approach not only encourages a greater number of responses, but minimizes the likelihood of stale groupthink.

这种做法不仅鼓励了更多的反应，而且减少跟风的可能性。