Description: 利用最大匹配法进行汉语句子的分词 最大匹配算法是最常用的分词算法，简单实用正确率可达到80%以上-the maximum matching method for the Chinese Sentence Word maximum matching algorithm is the most commonly used word segmentation algorithm, simple and practical accuracy rate can reach more than 80% Platform: |
Size: 73728 |
Author:廖剑 |
Hits:
Description: 一个简单的中文分词perl脚本，准确率比中科院的那个组件差点，但是全源码公开，可以扩展-A simple perl script Chinese word segmentation accuracy rate than the Chinese Academy of Sciences of the components was almost, but the whole open-source, can be extended Platform: |
Size: 359424 |
Author:kmsdoc |
Hits:
Description: 基于最大匹配算法的的中文分词组件，达到很好的分词准确率-Based on the maximum matching algorithm of the Chinese word segmentation components, to achieve good segmentation accuracy Platform: |
Size: 31744 |
Author:大宝 |
Hits:
Description: 图像处理中的图像分割grabcut算法，经试验可以达到很高的精确度，比较难，值得研究-Image Processing Image Segmentation grabcut algorithm, the test may be achieving a high degree of accuracy, more difficult, is worth studying Platform: |
Size: 5355520 |
Author:石红伟 |
Hits:
Description: KTDictSeg 简介: KTDictSeg 是由KaiToo搜索开发的一款基于字典的简单中英文分词算法
* 主要功能: 中英文分词，未登录词识别,多元歧义自动识别,全角字符识别能力
* 主要性能指标:
* 分词准确度:90％以上(有待专家的权威评测)
* 处理速度: 600KBytes/s-KTDictSeg Profile: KTDictSeg english by KaiToo developed a simple dictionary based on English and Chinese word segmentation algorithm* Main function: Chinese and English word, unknown word recognition, automatic recognition of multi-ambiguity, full-width character recognition ability* The main performance indicators:* Segmentation accuracy: 90 (subject to the authority of the expert evaluation)* processing speed: 600KBytes/s Platform: |
Size: 516096 |
Author:tz1985 |
Hits:
Description: 15篇光流配准经典文献，目录如下：
1、A Local Approach for Robust Optical Flow Estimation under Varying
2、A New Method for Computing Optical Flow
3、Accuracy vs. Efficiency Trade-offs in Optical Flow Algorithms
4、all about direct methods
5、An Introduction to OpenCV and Optical Flow
6、Bayesian Real-time Optical Flow
7、Color Optical Flow
8、Computation of Smooth Optical Flow in a Feedback Connected Analog Network
9、Computing optical flow with physical models of brightness Variation
10、Dense estimation and object-based segmentation of the optical flow with robust techniques
11、Example Goal Standard methods Our solution Optical flow under
12、Exploiting Discontinuities in Optical Flow
13、Optical flow for Validating Medical Image Registration
14、Tutorial Computing 2D and 3D Optical Flow.pdf
15、The computation of optical flow
-15 light流配quasi-classical literature, the directory is as follows: 1, A Local Approach for Robust Optical Flow Estimation under Varying2, A New Method for Computing Optical Flow3, Accuracy vs. Efficiency Trade-offs in Optical Flow Algorithms4, all about direct methods5, An Introduction to OpenCV and Optical Flow6, Bayesian Real-time Optical Flow7, Color Optical Flow8, Computation of Smooth Optical Flow in a Feedback Connected Analog Network9, Computing optical flow with physical models of brightness Variation10, Dense estimation and object-based segmentation of the optical flow with robust techniques11, Example Goal Standard methods Our solution Optical flow under12, Exploiting Discontinuities in Optical Flow13, Optical flow for Validating Medical Image Registration14, Tutorial Computing 2D and 3D Optical Flow.pdf15, The computation of optical flow Platform: |
Size: 14508032 |
Author:zhangji |
Hits:
Description: 针对类圆形堆积物图像的前景和背景在色彩或灰度上相近，难以用传统阈值分割等算法进行有效分割的问题，提出一种多方法融
合的类圆形堆积物图像区域分割策略。对图像进行滤波等一系列预分割处理后进行投影得到目标图像的外接矩形区域，以排除噪声的干扰，
在区域内采用改进的灰度共生矩阵方法进行粗分割，以解决窗口大小与分割精度的问题，采用二维OTSU 阈值分割方法对粗分割结果进行
量化。实验结果表明，该方法得到的区域分割结果边缘清晰、准确度高。-Round deposits for the prospect of images and background color or grayscale at up close, it is difficult using traditional threshold segmentation algorithm, such as the effective partition of the problem, a method of integration of many types of deposits circular image region segmentation strategy. Image filtering deal with a series of pre-partition after the target image projection external rectangular area in order to rule out the possibility of noise interference, in the region to adopt improved methods of co-occurrence matrix gray coarse segmentation to solve the window size and segmentation accuracy problem, using two-dimensional OTSU threshold segmentation method to quantify the coarse segmentation results. Experimental results show that the method of region segmentation results of the edge of a clear, high accuracy. Platform: |
Size: 141312 |
Author:子登 |
Hits:
Description: 在GPU中实现了快速目标分割，是比较小巧的程序，不能保证分割的精度-GPU at goal in the implementation of the Express division, are relatively small, the procedure can not guarantee the accuracy of segmentation Platform: |
Size: 10240 |
Author:nokia |
Hits:
Description: 指南针边缘检测算法是当前质量最高的边缘检测方法，比CANNY算子效果还要好。具有边缘定位准确，边连续，完整的特点。程序能够得到每一个像素的边的朝向，有利于做进一步的优化应用，例如图像分割等。具体算法过程可参考斯坦福大学麦克博士发表的论文。主要思路就是以每个像素为中心，像罗盘一样转动边，一直找到两边差异最大的朝向。最后把所有的最优朝向连起来成为边缘。
代码用VC驱动，核心部分用matlab编写-Compass edge detection algorithm is currently the highest quality of edge detection methods, than CANNY operator even better results. With edge positioning accuracy, while continuous, full features. Program can be a pixel for each side towards is beneficial for further optimization of applications, such as image segmentation and so on. Specific algorithms may make reference to Stanford University Maikeboshi the published papers. Main idea is to each pixel as the center, the same rotation as the compass edge, has been to find the greatest difference between both sides of the orientation of. Finally all the best to link into the edge towards. Code using VC-driven, the core part of the preparation with matlab Platform: |
Size: 82944 |
Author:黄沛杰 |
Hits:
Description: 指纹图像的质量测量与评价,在指纹图像分割、增强及指纹匹配等环节都有重要应用. 同时,指纹图像的质量分类,对指纹识别算法的适用性研究也有重要意义. 本文提出一种基于支持向量机的指纹图像质量分类方法.该方法选择梯度、Gabor特征、方向对比度等指标,利用支持向量机有效实现指纹图像质量分类. 并采用少类样本合成过采样技术( SMOTE)降低指纹图像质量好坏的类别不平衡问题对分类的影响. 理论分析和实验结果都表明该方法能够较为有效地提高指纹图像质量分类的正确率.-Fingerprint Image Quality Measurement and evaluation, in the fingerprint image segmentation, enhancement and fingerprint matching has important applications such links. Meanwhile, the quality of fingerprint image classification, on the applicability of fingerprint recognition algorithm is also important. In this paper, based on support vector machine classification of fingerprint image quality. The method chosen gradient, Gabor features, such as indicators of the direction of contrast, the effective use of support vector machine classification of fingerprint image quality. and use a small sample of synthetic over-sampling technique class (SMOTE) lower quality fingerprint image the type of imbalance bad influence on the classification. Theoretical analysis and experimental results show that the method can be used effectively to improve the quality of fingerprint image classification accuracy. Platform: |
Size: 564224 |
Author:郭事业 |
Hits:
Description: 是设计并实现一个汉语自动分词系统。在分析了自动分词面临的主要困难和难点的基础上，旨在降低分词难度和提高分词精度，设计并实现了一个基于正向最大匹配的汉语自动分词系统。-The design and implementation of a Chinese word segmentation system. In the analysis of word segmentation and difficulty of the major difficulties facing based on the word to reduce the difficulty and improve the segmentation accuracy, design and implementation of a maximum matching based on positive Chinese word segmentation system. Platform: |
Size: 1024 |
Author:杨亮 |
Hits:
Description: cv模型 matlab 图像分割源码，与传统模型相比，该C V模型具有更高的分割精度。-cv model segmentation matlab source, compared to the traditional model, the CV model has higher accuracy of segmentation. Platform: |
Size: 18432 |
Author:SylnaHu |
Hits:
Description: 针对合成孔径雷达(SAR) 图像含有大量斑点噪声的特点,基于Contourlet 的多尺度、局部化、方向性和各向
异性等优点,并结合隐马尔科夫树( HMT) 模型和隐马尔科夫场(MRF) ,提出了一种基于Contourlet 域持续性和聚
集性的SAR 图像模糊融合分割算法。该算法有效捕获了Contourlet 子带的持续性和聚集性,并分别用HMT 和
MRF 来刻画,再依据模糊测度,将多尺度HMT 和MRF 有机融合,建立Contourlet 域HMT2MRF 融合模型,并导
出新模型下的最大后验概率(MAP) 分割公式。对实测SAR 图像进行了仿真,仿真结果和分析表明:与小波域上的
HMT2MRF 融合分割及Contourlet 域上HMT 和MRF 分割算法相比,该算法在抑制斑点噪声的同时,有效地提高
了SAR 图像的分割精度- In view of the speckle noise in the synthetic aperture radar (SAR) images , and based on the Contourlet′s
advantages of multiscale , localization , directionality , and anisot ropy , a new SAR image fusion segmentation
algorithm based on the pe rsis tence and clustering in the Contourlet domain is p roposed. The algorithm captures the
pe rsis tence and clus tering of the Contourlet t ransform , which is modeled by hidden Markov t ree (HMT) and Markov
random field (MRF) , respectively. Then , these two models are fused by fuzzy logic , resulting in a Contourlet
domain HMT2MRF fusion model . Finally , the maximum a poste rior (MAP) segmentation equation for the new fusion
model is deduced. The algorithm is used to emulate the real SAR images . Simulation results and analysis indicate that
the p roposed algorithm effectively reduces the influence of multiplicative speckle noise , imp roves the segmentation
accuracy and p rovides a bet te r visual quality for SAR images ove r the Platform: |
Size: 897024 |
Author:周二牛 |
Hits:
Description: 针对FCM算法的运行时间长和计算量大的问题，提出了改进的FCM算法，先将图像分割成窗口大小的子块，然后以子块为单位提取特征向量，用FCM聚类粗分割，然后对边缘子块，以像素为单位从新提取特征向量，进行细分割。分割后的结果提高了运行速度和分割精度。-For the FCM algorithm and the calculation of long run the problem of large proposed improved FCM algorithm, first image into blocks the size of the window, and then sub-block feature vector extraction unit, using FCM clustering coarse partition, and then block on the edge, in pixels from the new Feature Extraction, for fine segmentation. Improve the segmentation results after the speed and accuracy of segmentation. Platform: |
Size: 3078144 |
Author:殷皓君 |
Hits:
Search - segmentation accuracy