搜索资源列表
wavelift
- 暂时只支持jpeg2000支持的 cdf97 和spline53 可以这样来测试: x=imread( E:\\study\\jpeg2000\\images\\lena.tif ) % see the decomposition coefficients y=wavelift(x, 1, spl53 ) using spline 5/3 wavelet figure subplot(1,2,1) imshow(x) subplot(1,2,2) imshow(mat2g
HHHH
- Type Classification Code [英汉计算机大词典] n.分类码 :main.m (program control)discretize.m (converts image to discrete values)plotimg.m (plots划分 images)dirImg.m (computes the directional image)extract.m (extracts square portion of image抽取块方向图 - called by dirImg
fingerprint
- Type Classification Code: main.m (program control) discretize.m (converts image to discrete values) plotimg.m (plots images) dirImg.m (computes the directional image) extract.m (extracts square portion of image - called by di
234
- 本设计是针对第二题采用了两种算法用来识别各种水果。第一种算法是质心法,第二种方法是相关匹配法。 质心法 一.算法思想 质心法的优点在于图像经过任何旋转后,图像边缘到质心的距离求和始终为一个固定值,这样利用这一个特性,就可以辨识出各个图像。 -This design is the second title for the two algorithms used to identify a variety of fruits. The first is a qualitative-f
DiscreteCosineTransform
- DCT A discrete cosine transform (DCT) expresses a sequence of finitely many data points in terms of a sum of cosine functions oscillating at different frequencies. DCTs are important to numerous applications in science and engineering, from lossy com
matlab-4
- 图像融合的目的把来自多传感器的数据互补信息合并成一幅新的图像,以改善图像的质量。图像融合最简单的理解就是两个(或多个)图像间的相加运算。这一技术广泛应用于多频谱图像理解和医学图像处理等领域。主要分为空域和频域相加。-The purpose of image fusion from multiple sensor data to complement the information merged into a new image to improve image quality. The simp
motion_length
- 采用投影求和统计图像的整列信息,对应图像的频域变换图像的黑带,用Round(kN/a)自动搜索得到最小值点,简易的与左右相邻2个像素比较判断最小值点,将两最小值点相减即得任意角度的运动模糊长度-Sum projection images using statistical information on the entire column, corresponding to the frequency domain transform image image Black Belt, with th
image-operations-
- 3 图像基本处理操作(图像显示、读写、像素统计处理、图像文件I/O等) 4 图像质量的客观评价(峰值信噪比PSNR及其应用) 5 图像的代数运算(绝对值差函数imabsdiff,叠加函数imadd,图像的旋转imrotate) 6 视频处理(创建AVI视频) 7 图像增强(直方图、直方图均化、均值滤波、中值滤波) 8 图像形态学分析(膨胀、腐蚀、开、闭运算、轮廓提取) -3 basic image processing operations (image display,
Fusion_old
- A novel higher order singular value decomposition (HOSVD)- based image fusion algorithm is proposed. The key points are given as follows: 1) Since image fusion depends on local information of source images, the proposed algorithm picks out info
CV6
- 利用KLT跟踪算法进行兴趣点选取和跟踪。 KLT跟踪算法的原始思想是在研究不同图像之间的匹配问题时,通过计算两个平移窗口的灰度残差,并寻找最小化残差SSD(sum of square difference)来实现匹配的。但是这个过程是没有效率的,因此KLT算法进行了优化。在这个过程中,KLT算法使用泰勒展开直接计算平移矢量,而不需要通过遍历进行搜索。 -KLT tracking algorithm to select a point of interest and tracking. K
SDD
- Function-Compute Correlation between two images using the similarity measure of Sum of Squared Differences (SSD) with Right Image
dct
- A discrete cosine transform (DCT) expresses a finite sequence of data points in terms of a sum of cosine functions oscillating at different frequencies. DCTs are important to numerous applications in science and engineering, lossy compression of audi
4
- 将彩色图像或者灰度图片进行切割,然后把切割后的图片相加或者相减-The color image or gray images were cut, and then put the sum of the picture after cutting or subtraction
LBP
- (1)计算图像中每个像素点的LBP模式(等价模式,或者旋转不变+等价模式)。 (2)然后计算每个cell的LBP特征值直方图,然后对该直方图进行归一化处理(每个cell中,对于每个bin,h[i]/=sum,sum就是一副图像中所有等价类的个数)。 (3)最后将得到的每个cell的统计直方图进行连接成为一个特征向量,也就是整幅图的LBP纹理特征向量; 然后便可利用SVM或者其他机器学习算法进行分类识别了。((1) calculate the LBP pattern of each p
data_batch_2
- cifar-10数据集由10个类的60000个32x32彩色图像组成,每个类有6000个图像。有50000个训练图像和10000个测试图像。数据集分为五个训练批次和一个测试批次,每个批次有10000个图像。测试批次包含来自每个类别的恰好1000个随机选择的图像。训练批次以随机顺序包含剩余图像,但一些训练批次可能包含来自一个类别的图像比另一个更多。总体来说,五个训练集之和包含来自每个类的正好5000张图像。 具体:batch2.mat文件,该训练集可以用于图片识别,非负矩阵分解等。(The ci