3d-FDTD-UPML__3D

说明：  图像特征提取的四种算法，花了很大的力气才弄来的！和大家分享！(Four types of image feature extraction algorithm, to spend a great deal of effort to get the! To share with you all!)

用于matlab图像处理,计算向量均值C，协方差m。 [C,m]=covmatrix(X)(Calculate vector mean C, covariance M.)

说明：  采用haar小波实现二维图像的分解与重构（mallat算法）：(Two-dimensional image using haar wavelet decomposition and reconstruction (mallat algorithm):)

形状上下文算法用于图像检索，程序中已运用了.mat文件存储图像库中的图像特征，检索效果很好。(shapecontext is used to realize imgage retrival,it shows very effective performance.)

本段代码的功能是通过滤波器对细胞图像先进行预处理，然后再提取细胞中心实现细胞的追踪，准确率比较高。(This code is the function of cells through the filter of the first pre-process the image, then extract the cell center to achieve cell tracking, accuracy is relatively high.)

小波去噪,实现阈值小波去噪，并且含有frost去噪，并且输出信噪比(Wavelet denoising, to achieve the threshold wavelet denoising, denoising and contains frost, and the output signal to noise ratio)

新的水下图像对比度增强和分割算法（实现对比度的增强和图像分割，内容包括：新的水下图像对比度增强算法、水下图像的分割算法--2008a）(The new underwater image contrast enhancement and segmentation algorithm (implemented contrast enhancement and image segmentation, including: new underwater image contrast enhancement algorithms, underwater image segmentation algorithm- 2008a))

使用ROEWA算子对有噪声SAR图像进行边缘检测(Use ROEWA operator of noise SAR image edge detection)

提出一种基于小波变换的图像配准方法。该方法首先用小波变换提取边缘特征点, 同时构成小波金字塔影像。在最高层利 用角度相关系数确定基准图像和配准图像之间的旋转角度, 经多分辨分级搜索匹配得到最终匹配特征点对, 利用这些特征点对待 配准图像进行配准纠正。实验结果表明, 改进方法有较好的效果, 可用于广泛应用领域。(Proposed a wavelet-based image registration method. In this method, using wavelet transform to extract the edge feature points, the same time constitute wavelet pyramid image. Point at the highest level the correlation coefficient using the baseline image registration between image and the angle of rotation, the multi-resolution hierarchical searching and matching features of the final match point, the use of these features point towards registration image registration correction. Experimental results show that the improved method has better results, can be used for wide range of applications.)

人工选择聚类中心，用k-means聚类方法对图像进行分割，效果不错的(artificial selection cluster center, k-means clustering method for image segmentation, the good results)