Description: 本文主要研究的是偏微分方程在数字图像处理方面的应用。文章首先详细阐
述了尺度空间的概念和性质，并以热传导方程所确定高斯尺度空间为例给出了连
续尺度空间与偏微分方程的联系。接着文章由简单的线性扩散方程开始，介绍了
非线性同质扩散和非线性异质扩散，不仅给出了相应的具体扩散方程模型，而且
对它们在图像处理中的不同效果作了比较。-This paper studies the partial differential equations in the digital image processing applications. The article first elaborates the concept of scale space and nature, and to determine the heat conduction equation by the Gaussian scale space is given as an example a continuous scale space with the associated partial differential equations. Then a simple article from the beginning of the linear diffusion equation, introduce a non-linear homogeneous proliferation and the proliferation of non-linear heterogeneous, not only the corresponding specific diffusion equation model, but also for them to deal with different image effects were comparison. Platform: |
Size: 962560 |
Author: |
Hits:
Description: 包括图像分析的四部分代码：matlab扩散和高斯函数，线性扩散，线性复扩散，非线性扩散。-It contains four parts：
[1]MATLAB function:diffusion.m gauss.m
[2]Linear diffusion
Applying linear diffusion to images creating linear scale-space.
MATLAB code: demo_lin.m
Image: haifa1.bmp
[3]Linear complex diffusion
Applying linear complex diffusion creating Gaussian and Laplacian scale-spaces. MATLAB code: demo_cmplin.m
[4]Nonlinear diffusions:
Perona-Malik ["Anisotropic diffusion"]
Catte et al. regularization of P-M
Complex ramp-preserving diffusion
Nonlinear edge preserving diffusions.The classical Perona-Malik process: the value of the diffusion coefficient is reduced near edges estimated by the first derivative. Best applied to step edges. A regularized version by Catte et al- the gradient estimation for controlling the process is smoothed by a Gaussian. Proved to be mathematically well-posed. Ramp preserving complex diffusion- best for ramp-type edges. Results are smoother with almost no staircasing effects.
MATLAB code: demo_nldif.m
Image: ct_scan.bmp Platform: |
Size: 239616 |
Author:吴豪科 |
Hits:
Description: sift算子中高斯金字塔和DOG的生成，只是该算子的最初步骤-sift operator：the generation of Gaussian pyramid and DOG。it is only the first steps Platform: |
Size: 2048 |
Author:lili |
Hits:
Description: 本文件包含拉普拉斯-高斯算子 Unsharp Masking, 和Sobel 算子，并用上述算子处理Grey scale图像，文件里有测试图像。代码用visual studio 2008编写，可以直接运行。-This file contains Scale Space using Laplacian of Gaussian operator Unsharp Masking, and Sobel Operator ,and using these operators to handle grey scale images.Algorithms developed in the environment of visual 2008, can be run directly in visual 2008. Platform: |
Size: 4083712 |
Author:Jason |
Hits:
Description: this function returns the SIFT descriptors DESCR of the SIFT frames P defined on the octave G of the Gaussian scale space. Platform: |
Size: 5120 |
Author:mazoul82 |
Hits:
Description: DOG检测算子，能够构造高斯尺度空间，并检测其中的极值，从而得到有价值的特征点位置-DOG detection operator, the Gaussian scale space can be constructed, and tested one of the extremes, resulting in valuable feature position Platform: |
Size: 25600 |
Author:ll |
Hits:
Description: 高斯噪声的小波变换仍然是高斯分布的，它均匀分布在频率尺度空间的各部分，而信号则由于其带限制性，它的小波变换系数仅仅集中在频率尺度空间上的有限部分。-Wavelet transform of the Gaussian noise is still Gaussian distribution, it is uniformly distributed in various parts of the frequency scale space signal because its a restrictive, its wavelet transform coefficients only focused on the frequency scale space on a limited part. Platform: |
Size: 2048 |
Author:cxy |
Hits:
Description: 把图像放置在Gaussian Scale Space，并求出各层图像的Hessian 矩阵。 通过Hessian举证，来求出图像中的blob和ridge。-The image is placed in the Gaussian Scale Space, and calculate the Hessian matrix of the layers of the image. At last the blob and ridge of the image will be detected. Platform: |
Size: 4096 |
Author:Rui Xiong |
Hits:
Description: SIFT 由特征提取，特征描述符描述和特征匹配 3 部分构成，该算子特征提取数目庞大，建立特征描述符运算
量高，导致算法效率低。提出了一种 SEC( SIFT-Edge-Corner) 算法，在图像尺度空间提取角点代替 SIFT 特征点，并根
据角点是边缘曲率极值理论，预先采用 Canny 算子得到高斯边缘图像金字塔，再提取角点并进行尺度选择。实验结
果表明: 该算法在保障高准确率的前提下大幅度提高特征提取效率-By the SIFT feature extraction, feature descriptions and feature matching descriptors 3 parts, the large number of feature extraction operator established feature descriptor computation high, resulting in low efficiency of the algorithm. Presents a SEC (SIFT-Edge-Corner) algorithm, the image scale space instead of SIFT feature extraction corner points and corner points based on extreme value theory is an edge curvature in advance using Canny operator edge image obtained Gaussian pyramid, and then extract corner point and scale selection. Experimental results show that: the algorithm protect high accuracy under the premise of feature extraction efficiency greatly improved Platform: |
Size: 231424 |
Author:焦婷 |
Hits:
Description: SIFT算法大致有四个步骤:
1，尺度空间极值检测。在尺度空间通过高斯微分函数来检测潜在的对于尺度和旋转不变的兴趣点。
2，关键点定位。在兴趣点位置上，确定关键点的位置和尺度。3，方向确定。基于图像局部的梯度方向，给每个关键点分配方向。4，关键点描述。在每个关键点的领域内测量图像局部的梯度。最终用一个特征向量来表达。-SIFT algorithm roughly four steps:
1, the scale space extremum detection. In the scale space to detect potential points of interest for the scale and rotation invariant differential by a Gaussian function.
2, the key point positioning. In point of interest position, determine the location and scale of key points. 3, to determine the direction. Based on local gradient direction, the direction assigned to each key. 4, the key point description. Local image gradient measurements within each key field. With a final feature vector to express. Platform: |
Size: 376832 |
Author:梅兰竹菊 |
Hits:
Description: 为在图像处理、计算器视觉、信号处理上所使用的一项技术。 高斯金字塔本质上为信号的多尺度表示法，亦即将同一信号或图片多次的进行高斯模糊，并且向下取样， 藉以产生不同尺度下的多组信号或图片以进行后续的处理，例如在影像辨识上，可以藉由比对不同尺度下的图片，以防止要寻找的内容可能在图片上有不同的大小。 高斯金字塔的理论基础为尺度空间理论，而后续也衍生出了多分辨率分析。-
As a technique in image processing, computer vision, signal processing is used. Gaussian pyramid is essentially a multi-scale representation of the signal, that signal will be the same picture multiple times or Gaussian blur, and down sampling, in order to produce multiple sets of different scales for subsequent signal or image processing, for example in the image the identification, can be compared to the picture at different scales, in order to prevent the contents may have to look for a different size in the picture. The theoretical basis for the scale-space Gaussian pyramid theory, but also spawned a follow-up multi-resolution analysis. Platform: |
Size: 244736 |
Author:robin lee |
Hits:
Description: This a MATLAB implementation of the SIFT keypoint detector and descriptor -do_gaussian: generate Gaussian scale space of input image
do_diffofg: generate Difference of Gaussian (DoG) scale space
do_localmax: local extrema as the potential keypoints
do_extrefine: refine the keypoints by discarding the ones with low contrast and along an edge
do_orientation: compute the orientation of a support region of keypoint
do_descriptor: compute the descriptor of a keypoint based on image gradients.
do_match: match two images based on the nearest neighbor principle and spatial consistency.
do_sift: generate the SIFT descriptors for a given input image. It basically s all the functions above.
Platform: |
Size: 1257472 |
Author:崔雪红 |
Hits:
Description: SIFT特征点检测配准方法建立高斯尺度空间的程序，比较有用-Registration SIFT feature point detection method to establish the Gaussian scale space program, more useful Platform: |
Size: 2048 |
Author:Grace |
Hits:
Description: SIFT算法是在不同的尺度空间上查找关键点,而尺度空间的获取需要使用高斯模糊来实现,Lindeberg等人已证明高斯卷积核是实现尺度变换的唯一变换核-SIFT algorithm is to find the key points on the different scales of space and capture scale space requires the use of a Gaussian blur to achieve, Lindeberg, who has proved a Gaussian convolution kernel is the only transform kernel to achieve scale transformation Platform: |
Size: 948224 |
Author:王禄平 |
Hits:
Search - gaussian scale space