[模式识别(视觉/语音等)] a1
这是一个针对于车牌识别、车牌字符分割和车牌字符识别的算法，源代码完整。
This is a for license plate recognition, license plate character segmentation and license plate character recognition algorithm, the source code integrity.
(2015-08-02, Visual C++, 2328KB, 下载4次)
[模式识别(视觉/语音等)] renlianshibie
人脸识别正确率97.18 ,1秒内完成。
能够在复杂的光线和背景下依然正确工作，并支持多张人脸在一个画面上的检测，识别，验证。戴眼镜和黑色人种也能检测和识别。能从多角度检测到人脸。
并带人鼻，嘴，眼镜,人脸左右边界的定位功能。
本软件仅限用于可见光(如室内光,日光灯,节能灯,白炽灯),并要求用户主动配合采模板,主动配合识别.最好有USB摄像头
阳光人脸识别开发包(SDK)兼容性信息
1.人脸识别SDK平台：Windows，源码虽用VC编写，但未用GDI及第三方图库,故可易于移值到LINUX。
2.人脸识别SDK界面：SunLightFace.DLL,标准Windows API调用，附详细说明书。
3.人脸识别SDK支持语言。VC,VB,Delphi,C#。
(Delphi在调用函数时，传参和接收输出要做转换，可用VC做一个dll来测试Delphi传入的参数显出来是否正确。)
4.人脸识别SDK版本分类：主动识别可见光版（即本网页所介绍），被动识别可见光版，主动识别红外版。
5.作者能为SDK客户提供除SDK技术支持以外的人脸识别软件二次开发技术支持。
6.（主动识别可见光版）在添加正上方节能灯光源及加入镜头滤光片，并配以低照度高画质的摄像头时，将取得最佳效果 (2015-03-30, Visual C++, 1073KB, 下载24次)
[模式识别(视觉/语音等)] handdetection
通过单目摄像头实现手势跟踪并且可以识别四种手势实现智能控制，需要OPENCV0.97以上版本。
Achieved by single camera head and hand tracking can identify four gesture intelligent control, you need to OPENCV0.97 above. (2010-10-08, Visual C++, 1371KB, 下载151次)
[模式识别(视觉/语音等)] getPDF2
本文提出了一种新的车辆许可证盘子识别,并在此基础上提出了一种自适应图像分割方法
In this paper, a new algorithm for vehicle license
plate identification is proposed, on the basis of a novel adaptive
image segmentation technique (Sliding Windows) in
conjunction with a character recognition Neural Network. The
algorithm was tested with 2820 natural scene gray level vehicle
images of different backgrounds and ambient illumination.
The camera focused on the plate, while the angle of view and
the distance from the vehicle varied according to the
experimental setup. The license plates properly segmented
were 2719 over 2820 input images (96.4 ). The Optical
Character Recognition (OCR) system is a two layer
Probabilistic Neural Network with topology 108-180-36, whose
performance reached 97.4 . The PNN was trained to identify
multi-font alphanumeric characters from car license plates
based on data obtained from algorithmic image processing. (2010-04-11, Visual C++, 812KB, 下载11次)
[模式识别(视觉/语音等)] FreeICTCLAS
汉语词法分析系统ICTCLAS(Institute of Computing Technology, Chinese Lexical Analysis System)，该系统的功能有：中文分词；词性标注；未登录词识别。分词正确率高达97.58 (973专家组评测)，未登录词识别召回率均高于90 ，其中中国人名的识别召回率接近98 处理速度为31.5Kbytes/s。ICTCLAS的特色还在于：可以根据需要输出多个高概率结果，有多种输出格式，支持北大词性标注集，973专家组给出的词性标注集合。该系统得到了专家的好评，并有多篇论文在国内外发表。
Chinese Lexical Analysis System ICTCLAS (Institute of Computing Technology, Chinese Lexical Analysis System), the system s features are: Chinese word segmentation-of-speech tagging unknown word identification. The correct segmentation rate of 97.58 (973 expert group evaluation), identification of unknown word recall rate were higher than 90 , one of the Chinese names to identify the recall rate of nearly 98 rate of treatment 31.5Kbytes/s. Also features ICTCLAS is: can output a number of high probability of a result, have a variety of output formats, to support the Beijing University of-speech tagging sets, 973 expert group give a collection of-speech tagging. The system has been praised by experts, and a number of papers published at home and abroad. (2009-03-03, Visual C++, 4331KB, 下载57次)
[模式识别(视觉/语音等)] 文字识别程序
目前该手写体识别系统主要分为 预处理模块： 主要包括训练数据和识别数据的读取，归一化，二值化 特征提取模块：主要包括笔划方向特征和网格密度特征，还可以根据对识别率的要求继续增加其他特征 识别（分类器）模块：主要包括SVM方法和BP神经网络的方法，其中SVM方法的识别率较高，SVM＋网格密度特征， 在小字符集情况下，达到了识别率97％以上 采用OO思想编写，适合做二次开发
currently the handwriting recognition system consists of pretreatment modules : including training data and identification data read, in one, two values of feature extraction module : the main characteristics of the direction including strokes and grid density characteristics, but also based on the identification of the requirements to continue to increase other identification ( Categories) modules : include SVM methods and BP networks, which SVM method for the identification of high SVM+ mesh density characteristics, in a small set of characters to achieve a recognition rate of over 97% using OO ideological preparation, suitable for secondary development (2005-07-18, Visual C++, 330KB, 下载1741次)