Three dimensional vascular reconstruction

基于matlab的自动报靶系统，文件内包括matlab源码以及GUI图形界面，以及模拟射击的靶纸图像，程序运行无误，靶纸环数识别准确。(Automatic target scoring system based on MATLAB)

用VC++视觉检测程序，可以打开PCB板的照片文件，对其进行编辑、点运算、几何变换、正交变换、图像增强、形态变换、边缘与轮 廓检测、图像分析与复原、图像编码等功能。其中点运算包括图像反色、线性变换、阀值变换等功能；边缘检测包括Robert算子、 Sobel算子、Gauss-Laplace算子等；正交变换包括傅里叶变换、离散余弦变换、沃尔什变换；图像编码包括哈弗曼编码表、香农弗 曼编码表、行程编码、LZW编码等。 对于研究电子封装技术，视觉检测及电机精确控制的同学学习参考使用。(visual detecting for PCB.rar)

SIFT图像配准，已测试比较好用，请大家放心使用(SIFT image registration)

对于图像运动补偿的编码，老师给的源程序，很好用(Motion compensation for image coding, the teacher to the source, easy to use)

图像检测的算法，Faster R-CNN算法，先对整张图像进行卷积计算，然后通过感兴趣区域池化层(RoI Pooling Layer)将选择性搜索算法推荐出来的候选区域和卷积网络计算出的特征映射图进行融合，得到候选区域对应的特征矢量，这种共享卷积计算的操作极大地减少了卷积计算的次数。而且这些特征矢量的维度统一，方便后续的分类工作。通过感兴趣区域池化层处理卷积特征，并将得到的特征送往两个并行计算任务进行训练，分类和定位回归。通过这些方法和改进的框架，Fast R-CNN 用更短的训练和测试时长，取得了比 R-CNN 更好的效果(Faster R-CNN algorithm first convolutes the whole image, then fuses the candidate regions recommended by the selective search algorithm and the feature mapping maps calculated by the convolution network through the RoI Pooling Layer to get the corresponding feature vectors of the candidate regions, which greatly reduces the number of convolution calculations. Moreover, the dimension of these feature vectors is unified, which facilitates the subsequent classification work. The convolution feature is processed by the pooling layer of the region of interest, and the obtained feature is sent to two parallel computing tasks for training, classification and positioning regression. Through these methods and improved framework, Fast R-CNN uses shorter training and testing time and achieves better results than R-CNN.)

题目： 八宝粥的生产过程中需要用到大量的红豆（红蚕豆），而实际上从红豆产地购买得到的红豆中会参杂一些杂物和不合格红豆。为了保证八宝粥的质量，需要对红豆进行筛选，去除不合格红豆和其它杂物。目前国内在红豆筛选工序上仍然使用人工筛选的方式，存在人工成本高、精确度差、劳动强度大等缺点。因此八宝粥生产厂商迫切需求开发一款能够自动实现红豆筛选的自动化筛选机。针对上述需求，请设计筛选机中的基于可见光图像的红豆筛选算法，该算法应尽可能的准确的找到参杂在红豆原料中的不合格红豆和杂物。 附件包括两个不同实验环境下拍摄得到的红豆相关图片样本库，实验环境分别为自然光照条件和灯光条件。每个条件下的样本库包括合格、不合格、杂物和混合四类样本的图像，其中合格、不合格以及杂物的图像可用来训练或建模，混合类图像用来测试算法的准确性和实时性。同学可在自然光照环境下的样本库或者灯光条件下的样本库中任选其一，用于解决上述问题。 作业提交格式包括以下四个内容： 一、 问题分析 二、 算法设计 三、 实验结果(将自己选择的样本库中的混合样本图片中的样本个体进行分类，并给出可视化的分类结果。) (Title:Use a lot of red bean production process of eight-treasure porridge ( red bean ), but in fact from the Red areas available in red bean can blend some sundries and unqualified red bean. In order to ensure the quality of eight-treasure porridge, the need for screening of red bean, red bean and other debris removal is not qualified. At present in the domestic bean screening process still use manual screening method, has the disadvantages of high labor costs, poor accuracy, high labor intensity. Therefore, eight-treasure porridge manufacturers need to develop a can achieve automatic bean screening automatic screening machine. According to the above requirements, please design screening machine selection algorithm based on visible light image of the red, the algorithm should be as accurate as possible to find the mixed in the bean raw materials in the unqualified red bean and debris.Accessories include captured in two different experimental environment of the red bean image sample dat)

区域生长算法分割脑部图像， 两个种子点，解决完整分割白质的算法(Region growing algorithm for segmentation of brain images, the two seed points, solve the integrity of white matter segmentation algorithm)

基于MATLAB GUI的图像处理界面，包括基本的图像处理功能(Graphical interface construction based on image processing)

归一化鬼成像计算方法，比传统鬼成像算法更快更好(Normalized ghost imaging calculation method, faster and better than traditional ghost imaging algorithm)

一般自然图像的梯度分布符合重尾分布，重尾分布也就是超拉普拉斯分布，根据这个特点进行图像复原。(The gradient of a natural image in line with heavy-tailed distribution, which is ultra-heavy-tailed distribution Laplace distribution, image restoration based on this feature.)