搜索资源列表
自己关于TVL1方法的改进
- 自己关于TVL1方法的改进 用来去除人脸上面的光照得到人脸的纹理图像，对于光照下人脸识别有相当大的作用。也可以用来定位,Their own methods to improve on TVL1 people face to face to remove the light to be Face of texture images, for face recognition under illumination a significant role. Can also be used to pos
IamSeg
- 基于形态学商图像的光照归一化算法.复杂光照条件下的人脸／P,~J1是一个困难但需迫切解决的问题，为此提出了一种有效的光照归一化算法． 该方法根据面部光照特点，基于数学形态学和商图像技术对各种光照条件下的人脸图像进行归一化处理，并且将它 发展到动态地估计光照强度，进一步增强消除光照和保留特征的效果．与传统的技术相比，该方法无须训练数据集以 及假定光源位置，并且每人只需一幅注册图像，在耶鲁人脸图像库B上的测试表明，该算法以较小的计算代价取得了 优良的识别性能．-Face recogn
Equalization
- MATLAB实现人脸识别,光照归一化算法-MATLAB realization of face recognition, illumination normalization algorithm
illumination-normalization
- 这个c#编写的程序，用来对人脸图像进行预处理，从而提升人脸识别算法的性能。这里提出了3种用于人脸识别的图像预处理的光照归一化算法，即：Multiscale retinex和anisotropic 和isotropic平滑方法。-The c# Preparation procedures used for face image preprocessing, so as to enhance the performance of face recognition algorithms. Here p
FaceFeaDetectionDLL
- 本人自己开发的一个人脸检测程序，实现人脸的精确检测，可检测人脸的轮廓，有效去除了背景干扰，对环境依赖行小，希望对做人脸检测的朋友们有所帮助是Dll的-this source code is about face detection which is developed by myself, it achieve the accurately face detection. it can detect the contour of face, and can remove the backgrou
Surveyonhumaneyesdetectioninimages
- 对在已知人脸区域中检测人眼的方法进行综述, 将其分为常光源下的人眼检测和特种光源的人眼检测 两大类。较为全面地综述了当前国内外相关眼睛检测以及定位技术, 分析了各种方法的优缺点, 探讨了当前困 扰人眼检测技术的难点以及人眼检测技术的发展趋势。-This paper supplied a survey of the eyes detection in the area of the known face images by dividing theminto normal illumi
AnisotropiqueSmoothing
- source code for light normalisation presented by Ralph Gross in this paper @inproceedings{RGross_AVBPA_2003, author = "Ralph Gross and Vladimir Brajovic", title = "An Image Preprocessing Algorithm for Illumination Invariant Face Recognition",
IluminationNormalization4FaceRecognition
- this document shows how to normalize illumination and reducing the negative effects of illumination on face recognition.
MahSom_v1
- This code is the implementation of Mahalnobis SOM algorithm published in this article. Face recognition under varying illumination using Mahalanobis self-organizing map S Aly, N Tsuruta, RI Taniguchi - Artificial Life and Robotics, 2008 - Springe
A_New_Algorithm_ofFastFaceDetection_underComplexCo
- 使用一种新的假面变换和模板匹配的人脸检测算法，能够对复杂环境中大小不等的人脸 进行快速、准确的检测．算法首先通过假面变换来预测人脸位置上边沿的中心，然后对图像中的预 测位置进行模板匹配，设计了一类对照明变化、噪声干扰具有较强适应性的模板匹配方法，最后对 匹配结果进行验证，确定人脸准确位置．采用多种环境下的大量图片进行实验，结果显示该算法具 有较快的检测速度和较高的准确性及鲁棒性． -A novelalgorithm offacedetection based on mask
lpp
- 基于LPP的人脸识别模块，运用matlab7.0编写，识别率达到70 以上，能够很好的识别不同姿势，光照，表情的变化-LPP-based face recognition module, using matlab7.0 writing, over 70 recognition rate can be a very good identification of the different positions, illumination, expression changes in
illumination_norm
- 毕设时写的程序，主要是人脸识别中的光照处理方法，包括直方图均衡，对数变换，SQI，MQI，SI等。本程序基于opencv实现。-This program demonstrates some illumination normalization methods used in face recognition.Histogram equaliztion,Logarithm transform,SQI,MQI are included.This program is based on opencv.
detection2
- 分析了Retinex 理论的本质意义, 得出Retinex 输出图像本质上是相对反射率, 而相对反射率对光照不敏感, 从而将其应 用于光照情况复杂的人脸图像预处理-Analysis of the Retinex theory of the nature of meaning, derived Retinex output image is essentially a relative reflectance rate, while the relative reflectivity is n
INface_toolbox_illumination_invariant
- INface toolbox for illumination invariant face recognition
face_detection
- computer vision, automatic detection, human face, face candidates search, skin-colour determination, 2D colour space, 3D colour space, illumination independence.
Facerecognitionbasedonilluminationinvariant
- 基于光照不变量的人脸识别-Face Recognition,Based on Illumination Invariant
imag_improve_rgb
- This a two stage method in which at first image RGB compensated and then converted to YCbCr to normalise overall illumination of image.Its cascaded implementation of section 2.A of paper "A FAST SKIN REGION DETECTOR" by Phil Chen, Dr.Christos Greecos
PreFace
- 采用光照归一化算法，可以用于人脸识别的预处理。-Illumination normalization algorithm for face recognition.
FACE-RECOGNITION
- 此文的目的有三个：第一，当地连续均值量化变换特征是提出照明和传感器敏感操作在目标识别上。其次，注册稀疏Winnows网络分割，提出了加快原分类。最后，特点和分类相结合对于正面人脸检测任务。检测结果列 为MIT + CMU系统和BioID数据库。关于这人脸检测器，接收器操作特征曲线BioID数据库产生最好的结果公布。对于结果麻省理工学院的中央结算系统+数据库相当于国家的最先进的脸探测器。一个人脸检测算法的MATLAB版本可以从http://www.mathworks.com/matlabce
illumination
- 光照不变的人脸识别问题是目前比较关注的一个大问题，本文章为你提供了一种新方法-a good paper for illumination face recognition method