资源列表
Haar_like
- Haar-like特征遍历，并求取不同特征在24*24图片下的特征个数-Haar-like feature traversal, and obtains the number of different features in the feature image of 24* 24
image-inpainting
- criminisi基于样本的图像修复算法，可以直接在VS上运行-criminisi inpainting
gray
- 全变分方法应用于恢复一幅退化图像。一个非常原始的方法。-The TV method is applied to recover a degraded image. A very primitive wmethod.
libsvm
- 用libsvm来实现图像分割，测试图片用的亦是25cases和40cases中的那个littleduck测试图片。主体程序思想为25cases中的代码过程，改进之处为可以让用户利用ginput来提取背景的样本点和前景（待分割出来的目标）的样本点作为训练样本，而不需实现指定背景和前景的样本点，也不用额外的小软件来查看某点的RGB值，ginput即可。-With libsvm to image segmentation, test picture is also used in 25cases an
Marr
- 利用MARR算子进行边缘检测，适用于图像的边缘提取-MARR operator using edge detection, the edge extraction for images
levelset
- 基于水平集的图像分割算法，用于对数字图像的分割，对于不复杂图像非常的精确。-Image segmentation algorithm based on level set, used for digital image segmentation, for complex image is very accurate.
fattal
- fattal图像去雾算法，有实验图像，能运行出结果，在前人的基础上改进的算法。-fattal image defogging algorithm, experimental images, run results, improved on the basis of the previous algorithm.
improved-algorithm
- 同态滤波图像去雾算法，有实验图像，能运行出结果，在前人的基础上改进的算法。研究生阶段研究的，希望对研究图像去雾的学者有帮助。-Homomorphic filtering algorithms of image to fog, a experimental images, can run out as a result, on the basis of predecessors improved algorithm. Graduate level research, hoping to help
Time-Frequency-Toolbox
- 时频分析工具箱，各种变换，图像处理-the Time-Frequency Toolbox
region_growing
- 图像处理算法 -基于区域生长的图像分割算法，对于不是很复杂的图像，处理精度很高。-Image processing algorithms the image segmentation algorithm based on region growinig, for the image is not very complex, high processing precision.
EM
- EM算法，能够比较准确的实现对二维数据的分类，已测试通过，完全可以运行-The realization of the EM algorithm, can be more accurate to 2 d data classification, tested through, can run
picture_process
- 对图像进行均衡处理的代码实现。能够对图像灰度值进行分析处理，得到图像灰度直方图，滤除两边的频率分量，保留中间的频率。-Code image equalization process implemented. Possible image intensity analysis processing to obtain an image histogram, filter out frequency components on both sides, to retain the intermedia