资源列表
splitBregmanROF_mex
- 分裂Bregman迭代法是将非线性先验分裂成 范数的和，用于提高算法的收敛速度，获得模型的最优解-Split Bregman iterative method is split into a nonlinear transcendental norm and, for improving the convergence rate, access to model the optimal solution
Natural-Image-Matting
- 影像抠图的目的是利用有限的样本，从影像背景中提取前景目标，并估计前景的透明度，即前景掩膜-The purpose is to use image matting limited sample, extract foreground objects the background image, and estimate the prospects of transparency, that the prospects mask
Chan-Vese-model
- Chan-Vese模型是基于区域的几何活动轮廓模型,此模型根据灰度级的同质性进行分类，图像域包括背景区域与前景区域。-Chan-Vese geometric model is region-based active contour model, this model are classified according to the gray level of homogeneity, including the background image field area and the foregro
sblb
- 经典的双边滤波程序，可以用于图像去雾处理，具有较好的实用性。-Classic bilateral filtering process, can be used for image processing, has good practicability.
test_gray_gradient
- 将一幅图像划分成8*8图像块，计算灰度梯度共生矩阵，基于混合熵对每个图像块分类-divide an image into 8*8 image blocks,calculate gray-gradient-matrix, classify image blocks based on the mixed entropy
gaijin_hessian
- 一种改进后的海森矩阵算法代码，针对医学图像处理效果比较好，可以识别较细的血管。-An improved Hessian matrix algorithm code, medical image processing effects for better, you can identify smaller vessels.
TImageColorspace
- 各种颜色空间的转化 史上最全的 有RGB HSI HSV Ycbcr LAB -The history of the various color space conversion has the most complete RGB HSI HSV Ycbcr LAB
xiangjian_yuzhi
- 针对相似图像，提取其不同的部分。采用先相减再进行阈值判断。-For similar images extracted different parts. Then subtract using the first threshold value judgment.
xiangjian_hessian
- 先进行图片相减，再进行海森变换，可提取图片中的变化区域。-First image subtraction performed, then Hassan transformation can extract picture change area.
1phog
- 学习层次加权再金字塔特征提取与分类，主要实现直方图特征提取与金字塔多层特征提取-learn the level weights when merging all the pyramid levels
1DeepLearnToolbox-master
- 深度学习matlab工具箱，已经测试通过的源码，可以应用在图像识别、模式识别、语音识别等领域-Deep learning matlab source code, which can be applied in graph identification, pattern recognition and sound recognition
1forest
- 森林算法实现，主要实现多特征分类实现，代码通过实际测试，可以实现分类任务-forest algorithm can be used to implement multi-features classification task