Description: 现有的代数特征的抽取方法绝大多数采用一维的方法，即首先将图像转换为一维向量，再用主分量分析（PCA）,Fisher线性鉴别分析（LDA）,Fisherfaces式核主分量分析（KPCA）等方法抽取特征，然后用适合的分类器分类。针对一维方法维数过高，计算量大，协方差矩阵常常是奇异矩阵等不足，提出了二维的图像特征抽取方法，计算量小，协方差矩阵一般是可逆的，且识别率较高。-existing algebra feature extraction method using a majority of the peacekeepers, First images will be converted into one-dimensional vector, and then principal component analysis (PCA), Fisher Linear Discriminant Analysis (LDA), Fisherfaces audits principal component analysis (KPCA), and other selected characteristics, then use the appropriate classification for classification. Victoria against an excessive dimension method, calculation, covariance matrix is often inadequate singular matrix, a two-dimensional image feature extraction method, a small amount of covariance matrix is usually reversible, and the recognition rate higher. Platform: |
Size: 2513 |
Author:小弟 |
Hits:
Description: 本MATLAB程序 用MATLAB 来实现3D 图像的匹配 用高斯卷积 来寻找 关键点 ,再利用关键点处的特征向量 通过3D向量的匹配 来实现3D 图像的配准-The MATLAB program using MATLAB to realize the matching of 3D images using Gaussian convolution to find the key points, re-use the key points of the feature vector through the 3D vector matching to realize 3D image registration Platform: |
Size: 1024 |
Author: |
Hits:
Description: 提出一种用于车辆牌照定位的新方法。该方法利用遗传算法对图像进行优化搜索,结合区域特征矢量构造的适
应度函数,最终寻找到牌照区域的最佳定位参量。实验结果表明,该方法抗噪声的能力强,提取出的牌照准确、完
整,具有很好的实用价值。
-A vehicle license for new ways of positioning. This method is the use of genetic algorithm to optimize the image search, combined with regional tectonic feature vector fitness function, and ultimately to find the best positioning license regional parameters. Experimental results show that the method of anti-noise ability, to extract the license is accurate, complete, has good practical value. Platform: |
Size: 52224 |
Author:zoubinbin |
Hits:
Description:
利用马尔可夫模型估计的条件概率做为空间特征向量，用于图像检索。-Markov model using the estimated conditional probability space as a feature vector for image retrieval. Platform: |
Size: 1024 |
Author:Owenli |
Hits:
Description: 一种基于综合特征向量的图像检索算法（一种基于综合特征向量的图像检索算法）-A feature vector based on an integrated image retrieval algorithm (a feature vector based on an integrated image retrieval algorithm) Platform: |
Size: 182272 |
Author:gs |
Hits:
Description: 为进一步进行纹理特征分析，从纹理的方向性人手，给出了纹理方向的数学定义式，合理选择差异函数，
构造了具有物理意义的纹理方向描述特征向量．数据处理方面，运用模糊贴近度的概念，结合改进后的属性均值聚
类算法，对一类具有方向性的纹理图象进行分类与分割实验，取得了较好的结果．试验表明，该方法对纹理的方向
性有很好的描述能力．
关键词 图象分割 纹理方向 纹理分割 神经网络 模糊聚类
-Texture features for further analysis, staff from the texture direction, given the direction of the mathematical definition of texture type, a reasonable choice of the difference function, structure with a physical meaning to describe the direction of the texture feature vector. Data processing, the use of the concept of fuzzy nearness degree, combined with improved properties means clustering algorithm for a class of directional texture image classification and segmentation experiments, achieved good results. Tests show that the method of directional texture has a good description of the capacity. Keywords Image segmentation, texture segmentation texture direction fuzzy clustering neural network Platform: |
Size: 291840 |
Author:wgn |
Hits:
Description:
采用Gabor_Palm函数提取掌纹图像的能量特征，并将得到的结果分块，分别计算每块的均值和方差作为特征向量。特征向量的长度为160.-Gabor_Palm function using the energy extracted palmprint image features, and will be the result of sub-blocks, each block were calculated the mean and variance as a feature vector. Feature vector length is 160. Platform: |
Size: 2048 |
Author:wangxuyang |
Hits:
Description:
在特征提取阶段，研究了PCA, 2DPCA, (2D) 2PCA, DiagPCA, DiagPCA-F-2DPCA等多
种方法。不同于基于图象向量的PCA特征提取，由于2DPCA, (2D) ZPCA, DiagPCA和
DiagPCA-I-2DPCA的特征提取都直接基于图象矩阵，计算量小，所以特征的提取速度明
显高于PCA方法。-In the feature extraction stage, the study of the PCA, 2DPCA, (2D) 2PCA, DiagPCA, DiagPCA-F-2DPCA and other methods. Vector is different from the PCA-based image feature extraction, as 2DPCA, (2D) ZPCA, DiagPCA and DiagPCA-I-2DPCA the feature extraction are directly based on image matrix, a small amount of calculation, so the speed of feature extraction method was significantly higher than PCA . Platform: |
Size: 45056 |
Author:付采 |
Hits:
Description: 近年来，随着互联网的高速发展，网上的多媒体信息也急剧增加，这些多媒体信息以图像为主。如何从浩瀚的图像数据库中快速、准确地找出所需要的图像，己成为一个备受关注的研究课题。有效地组织、管理和检索大规模的图像数据成为迫切需要解决的问题。于是基于内容的图像检索(Content-Based Image Retrieval: CBIR)作为一个崭新的研究领域出现了。
本课题拟研究、分析彩色图像红、绿、蓝三基色直方图的生成、特征提取和相似度等问题，应用图像的颜色信息—三基色直方图对图像进行检索。针对基于颜色的图像检索，本文采用应用广泛的RGB颜色空间来表示图像的颜色特征，对颜色分量进行等间隔量化并形成特征矢量并对特征矢量进行归一化处理，采用图像均匀分块的方法引入图像中色彩所处的位置信息，用距离度量函数进行图像的相似性匹配。在此基础上实现了基于三基色直方图算法的检索系统。
本文的研究和实践对于促进基于内容的图像数据库检索技术的应用具有一定的参考价值和实践意义。-With the rapid development of Internet, the multimedia information is booming. All this information is mostly images. Effective recognizing, management and searching all these images have been an emergent problem. This has led the rise of a new research and development field: Content-Based Image Retrieval (CBIR).
The topics to research, analysis color images red, green and blue color histogram generation, feature extraction and the similarity of the issues, application of the color image- trichromatic histogram of the image retrieval. Based on the color against the static image retrieval, this paper application of a wide range of RGB color space to indicate the color image features, the color components, such as spacing and quantitative characteristics of a feature vector and a normalization of vector processing, using uniform image block the introduction of the method in which the color image Location information, and distance measuring function similar to the image of the match. On Platform: |
Size: 408576 |
Author:qichao |
Hits:
Description: The first scheme is based on the spatial locality of feature vectors corresponding
to similar images. Learning is effected by modifying the query vector to incorporate the
positive examples. The second scheme is based on “distorting” our view of the feature space. An
new similarity distance between an image and the query is learned from the relevance feedback. Platform: |
Size: 1024 |
Author:rach |
Hits:
Description: 对图像进行分主题过滤。
根据给定的图像样本集，进行样本训练，提取多个颜色空间中的多种方式(颜色矩，纹理谱，直方图，肤色模型等)的得到图像特征集；对待过滤图像进行特征提取，向量匹配，进而实现图像分主题分类功能。-Sub-theme of the image filter. According to a given set of image samples to conduct the training samples to extract more color space in multiple ways (color moments, texture spectrum, histogram, color model, etc.) are image feature set treatment filters for feature extraction, vector match, then the sub-theme classification of image features. Platform: |
Size: 9487360 |
Author:andy |
Hits:
Description: popfnn神经网络，它是一种模糊伪输出的神经网络(pseudo-outer-production fuzzy nerual network).可用于模式识别。运行pop1可以训练输入的特征向量，extractionpop使用来抽取图片的特征向量的。-popfnn neural network, it is a pseudo-output fuzzy neural network (pseudo-outer-production fuzzy nerual network). can be used for pattern recognition. Can be trained to run pop1 input feature vector, extractionpop used to extract the image feature vector. Platform: |
Size: 7168 |
Author:毛敏 |
Hits:
Description: 基于多帧图像插值（Interpolation）技术的方法是SR恢复技术当中最直观
的方法。这类方法首先估计各帧图像之间的相对运动信息，获得HR图像在非均
匀间距采样点上的象素值，接着通过非均匀插值得到HR栅格上的象素值，最后
采用图像恢复技术来去除模糊和降低噪声（运动估计!非均匀插值!去模糊和
噪声）。-In this paper, we propose a novel method for solv-
ing single-image super-resolution problems. Given a
low-resolution image as input, we recover its high-
resolution counterpart using a set of training exam-
ples. While this formulation resembles other learning-
based methods for super-resolution, our method has
been inspired by recent manifold learning methods, par-
ticularly locally linear embedding (LLE). Speci?cally,
small image patches in the low- and high-resolution
images form manifolds with similar local geometry in
two distinct feature spaces. As in LLE, local geometry
is characterized by how a feature vector correspond-
ing to a patch can be reconstructed by its neighbors
in the feature space. Besides using the training image
pairs to estimate the high-resolution embedding, we
also enforce local compatibility and smoothness con-
straints between patches in the target high-resolution
image through overlapping. Experiments show that our
method is very ?exible Platform: |
Size: 27595776 |
Author:qianyeyu |
Hits:
Description: 提出了一种可抗
几何攻击的二值文本图像水印算法. 该算法基于DCT,它将二值图像的DCT、图像的视觉特征向
量和加密技术有机结合起来,在提取水印时不需要原始图像,是一种实用的盲水印算法-We propose a geometric attacks against Binary Text Watermarking Algorithm. The algorithm is based on DCT, it binary image DCT, the image of the visual feature vector and encryption technology combine in the watermark without the original image is a practical blind watermarking algorithm Platform: |
Size: 1190912 |
Author:切尔西 |
Hits:
Description: 缘检测算子进行边缘检测；其次，利用Hu的7个不变矩作为形状特征向量；再次，进行图像的相识度匹配；最后在图像库中检索出最相近的Top10图像序列作为检索结果。-Edge detection operator edge detection Secondly, Hu' s seven moment invariants as the shape feature vector Again, a acquaintance of the image matching and finally in the image library to retrieve the most similar Top10 image sequence as a search result. Platform: |
Size: 4805632 |
Author:郑珊珊 |
Hits:
Description: 随着科学技术的飞速发展，机器学习与人工智能技术的不断创新，人们对特定信息检索的需求逐渐增加，使得如何对资源进行合理有效的分类成为一个关键问题。近几年来，基于内容的图像分类的研究焦点主要集中在自然图像的场景分类和物体分类两个方面，大多采用有监督学习方法，通过对底层特征建模和中间语义分析来实现分类。
本文基于Libsvm的图像分类研究及实现，主要针对的是物体分类这一方面，选用了五类水果作为分类研究的对象。对图像进行分类的大体步骤主要包括采集图像样本（主要从Web上获取）、图像预处理（如截成大小一致的图片）、特征向量提取、结合Libsvm进行模型训练、对测试图片进行分类测试。(With the rapid development of science and technology, the continuous innovation of machine learning and artificial intelligence technology, the demand of people for specific information retrieval increases gradually, and how to classify resources reasonably and efficiently becomes a key issue.
This article based on Libsvm image classification research and implementation, mainly for the object classification on the one hand, the selection of five types of fruit as the classification of the object of study. The general steps of image classification include collecting image samples (mainly obtained from the Web), image preprocessing (such as cutting the same size of the image), feature vector extraction, combined with Libsvm model training, the test images were classified test.) Platform: |
Size: 1024 |
Author:安安*
|
Hits:
Description: The feature vector of
an image can be derived from the histograms of its color
components and finally can set the number of bins in the color
histogram to obtain the feature vector of desired size. Thus the
grid code of an image is obtained through the quantization of the
feature vector derived from the histogram of the desired color
component of the image. In order to have similar features of the
images the grid code must be same for all Images in the grid Platform: |
Size: 2532942 |
Author:msujikumar87@gmail.com |
Hits:
Search - image feature vector