Description: Image combination. This program adds two image together from the row direction and automatically detect the size of the two images. Platform: |
Size: 1447 |
Author:张希勇 |
Hits:
Description: Image combination. This program adds two image together from the row direction and automatically detect the size of the two images. Platform: |
Size: 1024 |
Author:张希勇 |
Hits:
Description: ：给出了一种新的Hough 变换圆检测方法——点Hough 变换（PHT）。该方法根据圆周上任意两条不平行弦的中垂线相交与圆心的几
何性质，同时选取曲线上3点 进行计算，将传统Hough变换圆检测时的三维参量统计变成一维参量统计，极大地降低了计算复杂性和对资源的需求。为了克服任意选取组合点可能带来的计算量增加及统计结果的分散程度提高,文中提出了点的选择方法。合成图和实际图像的实验结果表明，该方法用于普通图像中圆检测时快速、稳定、准确。-: This paper presents a new method of Hough transform circle detection- point Hough transform (PHT). The method according to any circle on the two parallel strings do not intersect with the vertical center of the geometric properties, select the curve at the same time to 3:00, the traditional Hough transform circle detection parameters when the three-dimensional into a one-dimensional statistical parametric statistics, greatly reduces the computational complexity and demand for resources. In order to overcome the combination of any selected point of the calculation of the potential increase in the dispersion results and the degree of increase in the points raised in the text of the selection method. Synthesis of maps and images of the actual experimental results show that the method can be applied to general image circle detection in rapid, stable and accurate. Platform: |
Size: 212992 |
Author:王浩 |
Hits:
Description: 随着LIDAR技术的出现，三维建筑物的提取也受到越来越多的重视。由于LIDAR数据分布的不连续性和不规律
性，直接从机载激光扫描测距数据中进行建筑物提取较为困难。本文提出了一种结合灰度影像的LIDAR数据三维建筑物提
取方法，分三个步骤：首先对灰度影像进行建筑物二维提取；然后将处理后影像和LIDAR数据粗匹配，初步确定LIDAR数
据中的三维建筑物区域；最后利用一组阈值操作进行三维建筑物的精确提取。实验结果表明该方法简单实用，适应性强。-With the emergence of LIDAR technology, three-dimensional building extraction has also been more and more attention. LIDAR data distribution as a result of non-continuity and regularity, directly from the location of airborne laser scanning data extraction more difficult to carry out the building. In this paper, a combination of gray-scale image of the LIDAR data extraction method of three-dimensional buildings, divided into three steps: First of all buildings on the two-dimensional gray-scale image extraction then processed LIDAR data images and rough match, LIDAR data to determine the initial three-dimensional buildings in the region the final threshold of use of a group of operations extract accurate three-dimensional structures. The experimental results show that the method is simple and practical, adaptable. Platform: |
Size: 178176 |
Author:jerry |
Hits:
Description: 对图像的去噪处理，包括了加权梯度领域去噪、高斯模板去噪和两种方法的结合-Denoising of images, including the weighted gradient field denoising, Gaussian model for denoising and the combination of two methods Platform: |
Size: 1024 |
Author:zbj |
Hits:
Description: 汽车防撞，技术路径不外：1.雷达测距防撞；2.视差测距防撞。前者，一旦保有量较大，必定遭遇互相干扰问题；后者，以前主要问题是，算法复杂，实时性差。
本文公开了一种新算法（已申请发明专利），主要运算可以借助硬件组合逻辑模块并行执行，可以极大提高视差测距的实时性，满足汽车防撞的要求。-(Background) Stereo matching, requires in two images to identify two pixels to be matched each other, in the field of machine vision, this is a difficult, but central task To complete the task, the key is to strengthen pixels characteristics (status)In order to strengthen the characteristics of pixels , has developed combination of the pixel neighborhood features technology, a variety of window matching technology, and matching prominent features of pixels only, for example, the edge pixels in the image, or to match with the Sift technical characteristics of pixels The window matching Technology, in theory, the larger the window, the more prominent features of pixels to be matched, but the resulting: 1. the computational size is increase 2. the windows across the actual object boundary, so that， the same characteristics pixel in the other image is inexistence，making computing more difficult, so two questions The feature matching, can acquire sparse disparity map only, and mat Platform: |
Size: 318464 |
Author:李维纲 |
Hits:
Description: 采用一种基于极线约束与区域匹配相结合的立体匹配算法对两幅图像中对应点进行匹配。-Using the epipolar constraint and based on a combination of area-based matching algorithm for stereo matching corresponding points in two images to match. Platform: |
Size: 2234368 |
Author:崔小强 |
Hits:
Description: 局域网屏幕监控系统是主要由客户端和服务器端两部分组成。客户端模块该模块主要用于抓取屏幕信息，进行数据压缩，然后划分数据报，向服务器端发送数据，并等待服务器发来的确认信息。服务器端模块该模块主要用于接收收客户端发送的数据报，然后向客户端发送确认信息，接着组合数据报为JPEG数据流，最后显示JPEG图像。操作注意事项（1）可执行文件位置：Server\Debug\Server.exe Client\De
-The LAN screen monitoring system is mainly composed of two parts, the client and server side. The client module This module is used to capture screen information, data compression, and then divide the datagram to send data to the server-side, and wait for the server to confirm information. The server-side module The module is mainly used for receiving client received datagrams sent, then send a confirmation message to the client, followed by a combination of data reported to the JPEG data stream, and finally display JPEG images. The operation Note (1) executable file location: Server \ Debug \ Server.exe Client \ De Platform: |
Size: 2163712 |
Author:liming |
Hits:
Description: 运用matlab对图像进行放大：像素复制法和双线性插值法。像素复制方法的图像缩放的原理主要是对原来输入图像的整行或是整列像素进行简单的复制与删除，达到改变图像大小的目的。双线性插值放大算法中，目标图像中新创造的象素值，是由原图像位置在它附近的小区域象素的值通过加权平均计算得出的。-Write MATLAB function to zoom a grayscale image from original size to the given output
size through two different methods: a) pixel repetition and b) bilinear interpolation.
Zoom the images processed in the Task 2 back to the original size using above mentioned
methods.
Calculate MSE and PSNR for the original and zoomed images using MATLAB methods
written for the Task 1. Plot the MSE and PSNR as function of the size reduction ratio
respectively.
Define what combination of shrink/zoom methods provides the minimal MSE and maximal
PSNR for the same size reduction ratio. Platform: |
Size: 5120 |
Author:宁可 |
Hits:
Description: 本书系统地介绍了图像处理与识别的基本原理、典型方法和实用技术。全书共分12章，第1章～第6章是图像处理与识别的基础内容，包括图像科学综述、MATLAB语言图像编程、图像增强、图像分割、图像特征提取和图像识别；第7章～第10章是图像处理与识别的工程实例，涵盖了医学图像处理、文字识别和自导引小车路径识别等应用实例，并结合理论算法，提供了大量MATLAB代码程序，以帮助读者掌握如何使用MATLAB语言快速进行算法的仿真、调试和估计等方法。第11章～第12章，是两个综合性较强的实例，分别是Visual C++实现的基于神经网络的文字识别系统和车牌定位系统。 本书附带的光盘给出了各个章节列举的实例的源代码，同时赠送了28个常用数字图像处理算法的Visual C++代码实现。 本书讲解深入浅出，实例程序丰富，注重理论与实践相结合。本书可作为计算机应用、自动化、图像处理与模式识别、机电一体化专业的高年级本科生或研究生的参考书，也可供从事图像处理与识别的研究人员和工程技术人员阅读参考。-This book introduces the basic principles of image processing and recognition of the typical methods and practical skills . The book is divided into 12 chapters , Chapter 1- Chapter 6 is the basis of the content of image processing and recognition , including images scientific overview , MATLAB programming language images, image enhancement, image segmentation, image feature extraction and image recognition Chapter 7- Section Chapter 10 is a project example image processing and recognition , covering the medical image processing, character recognition and self- guided trolley path recognition example , the combination of theory and algorithms, for a lot of MATLAB code procedures to help readers learn how to use MATLAB language fast simulation , debugging and estimation methods algorithm. Chapter 11- Chapter 12 , are two examples of highly integrated , namely Visual C++ implementation based on neural network character recognition system and a license plate positioning system. The book Platform: |
Size: 6598656 |
Author:朱朴宁 |
Hits:
Description: 医学图像配准，医学图像2D_3D配准，基于vc++开发。-Abstract. We present a method that performs the rigid 2D/3D image registration
efficiently on the GPU. As one main contribution of this paper, we propose an
efficient method for generating realistic DRRs that are visually similar to X-ray
images. Therefore, we model some of the electronic post-processes of current X-
ray C-arm-systems. As another main contribution, the GPU is used to compute
eight intensity-based similarity measures between the DRR and the X-ray image
in parallel. A combination of these eight similarity measures is used as a new
similarity measure for the optimization. We evaluated the performance and the
precision of our 2D/3D image registration algorithm using two phantom models.
Compared to a CPU+GPU algorithm, which calculates the similarity measures
on the CPU, our GPU algorithm is between three and six times faster. In contrast
to single similarity measures, our new similarity measure achieved precise and
robust registration results for both phantom m Platform: |
Size: 329728 |
Author:刘坤 |
Hits:
Description: 局域网屏幕监控系统是主要由客户端和服务器端两部分组成。
客户端模块
该模块主要用于抓取屏幕信息，进行数据压缩，然后划分数据报，向服务器端发送数据，并等待服务器发来的确认信息。
服务器端模块
该模块主要用于接收客户端发送的数据报，然后向客户端发送确认信息，接着组合数据报为JPEG数据流，最后显示JPEG图像。
-LAN screen monitoring system is mainly composed of client and server-side two parts. client module The module is mainly used to grab screen information, data compression, and then divide the datagram to send data to the server, and wait for the confirmation message sent the server. server module The module is mainly used for receiving data packets sent by the client, and then send a confirmation message to the client, then a combination of the data reported for the JPEG data stream, and finally display JPEG images. Platform: |
Size: 17202176 |
Author:萧雨 |
Hits:
Description: DSmT证据推理的组合公式计算函数，表示出两帧图像间各个像素点的相对情况，用MATLAB实现动态聚类或迭代自组织数据分析。- Combination formula DSmT evidence reasoning calculation function, Between two images showing the relative circumstances of each pixel, Using MATLAB dynamic clustering or iterative self-organizing data analysis. Platform: |
Size: 4096 |
Author:李志宏 |
Hits:
Description: 压缩感知恢复算法，使用新的范数来提升图像恢复能力，包含论文和代码。-We propose a compressive sensing algorithm that exploits geometric properties of images to recover images of high quality few measurements. The image reconstruction is done by iterating the two following steps: 1) estimation of normal vectors of the image level curves and 2) reconstruction of an image fitting the normal vectors, the compressed sensing measurements and the sparsity constraint. The proposed technique can naturally extend to non local operators and graphs to exploit the repetitive nature of textured images in order to recover fine detail structures. In both cases, the problem is reduced to a series of convex minimization problems that can be efficiently solved with a combination of variable splitting and augmented Lagrangian methods, leading to fast and easy-to-code algorithms. Platform: |
Size: 1130496 |
Author:wf |
Hits:
Description: SeetaFace人脸识别引擎包括了搭建一套全自动人脸识别系统所需的三个核心模块，即：人脸检测模块SeetaFace Detection、面部特征点定位模块SeetaFace Alignment以及人脸特征提取与比对模块SeetaFace Identification。
主要功能：
人脸检测模块(SeetaFace Detection): 采用了一种结合传统人造特征与多层感知机（MLP）的级联结构，在FDDB上达到了84.4 的召回率（100个误检时），并可在单个i7 CPU上实时处理VGA分辨率的图像。
面部特征点定位模块(SeetaFace Alignment): 通过级联多个深度模型（栈式自编码网络）来回归5个关键特征点（两眼中心、鼻尖和两个嘴角）的位置，在AFLW数据库上达到state-of-the-art的精度，定位速度在单个i7 CPU上超过200fps。
人脸识别模块(SeetaFace Identification): 采用一个9层的卷积神经网络（CNN）来提取人脸特征，在LFW数据库上达到97.1 的精度（注：采用SeetaFace人脸检测和SeetaFace面部特征点定位作为前端进行全自动识别的情况下），特征提取速度为每图120ms（在单个i7 CPU上）。
-The SeetaFace Face Recognition Engine includes the three core modules required to build a fully automated face recognition system, namely the Face Detection Module SeetaFace Detection, the Face Feature Point Segment Module SeetaFace Alignment, and the Face Feature Extraction and Matching Module SeetaFace Identification.
The main function:
Face Detection Module (SeetaFace Detection): A combination of traditional artificial features and multi-layer sensor (MLP) cascade structure, in the FDDB reached 84.4 recall rate (100 false detection), and Can process VGA resolution images in real time on a single i7 CPU.
Face feature positioning module (SeetaFace Alignment): by cascading multiple depth model (stack self-coding network) to return to five key feature points (two centers, nose and two mouth) position, in the AFLW To achieve the state-of-the-art accuracy, positioning speed in a single i7 CPU more than 200fps.
Face Recognition Module (SeetaFace Identificat Platform: |
Size: 2048 |
Author:赵炳坤 |
Hits:
Description: 新的自下而上的视dmits combination with other maps. The model is simple, and biologically plausible
insofar as it is naturally parallelized. This model powerfully predicts human
xations on 749 variations of 108 natural images, achieving 98% of the ROC area
of a human-based control, whereas the classical algorithms of Itti & Koch ([2],
[3], [4]) achieve only 84%.觉显著性模型，基于视觉显著图（GBVS），提出了。(A new bottom-up visual saliency model, Graph-Based Visual Saliency (GBVS), is
proposed. It consists of two steps: rst forming activation maps on certain feature
channels, and then normalizing them in a way which highlights conspicuity and
admits combination with other maps. The model is simple, and biologically plausible
insofar as it is naturally parallelized. This model powerfully predicts human
xations on 749 variations of 108 natural images, achieving 98% of the ROC area
of a human-based control, whereas the classical algorithms of Itti & Koch ([2],
[3], [4]) achieve only 84%.) Platform: |
Size: 9145344 |
Author:还有叫二哈的
|
Hits:
Description: Between two images showing the relative circumstances of each pixel, Data packet transfer source program, Combination formula DSmT evidence reasoning calculation function. Platform: |
Size: 7168 |
Author:quipangang
|
Hits:
Description: halcon+delphi7的强悍组合。识别二维码。妥妥的。delphi7里面用halcon，资料非常少。本历程，包含了所有需要的二维码图片、halcon库函数、exe范例、pas源代码，非常珍贵的学习资料，全面展示了delphi里面用halcon的方法。(The intrepid combination of halcon+delphi7. Two-dimensional code identification. Properly. Delphi7, which uses Halcon, data is very small. This course contains all the required two-dimensional code images, Halcon library functions, EXE examples, PAS source code, very valuable learning materials, a comprehensive display of the Delphi inside, using the Halcon method.) Platform: |
Size: 993280 |
Author:John20170818
|
Hits:
Description: gbvs_nips 基于图的显著性。Graph-Based Visual Saliency。是发表在2006年NIPS上的论文。(A new bottom-up visual saliency model, Graph-Based Visual Saliency (GBVS), is proposed. It consists of two steps: first forming activation maps on certain feature channels, and then normalizing them in a way which highlights conspicuity and admits combination with other maps. The model is simple, and biologically plausible insorfar as it is naturally parallelized. This model powerfully predicts human fixations on 749 variations of 108 natural images, achieving 98% of the ROC area of a human-based control, whereas the classical algorithas of Itti & Koch ([2], [3], [4]) achieve only 84%.) Platform: |
Size: 8948736 |
Author:佳子 |
Hits:
Description: This paper presents a new algorithm for human action
recognition in videos. This algorithm is based on a combination of
two different feature types extracted from Aligned Motion Images
(AMIs). The AMI is a method for capturing the motion of all frames
in a human action video in one image. Platform: |
Size: 1661952 |
Author:nabb |
Hits:
Search - combination of two images