Realtime Face Recognition in Labview

author LokiXun [email protected]

FaceDetection Blog reference

• Labview 与 Matlab 接口实现 参考1 Matlab Script 的使用

usage

• Program Entry

  FaceRecognition_realtime_detection_usingPCA.vi
  

Database Preparation

• prepare person facial photos prepare each person's face image in the directory .\Face_Database\Name by person's name.

• generate faces' eigenvectors of each photo run .\FaceDB_Generate_Matlab\FaceRecognitionPCA_generate_DB_eigenvector.m to generate Face's eigenvector, and saved in mat file .\FaceDB_Generate_Matlab\FaceDetect_DB_23_2_9.mat

setup

• Labview setup
• Matlab version > 2018
• setup Labview's OpenCV Face Detection Module

Q&A

1. vision opencv库 dll文件找不到

现在 PC 上的 Labview 是 32 bit 的！！

2. 使用MATLAB®脚本节点时遇到错误1047

3. https://forums.ni.com/t5/kl3m3n-s-blog/Real-time-face-and-eye-detection-in-LabView-using-OpenCV-Harr/ba-p/3486952?profile.language=zh-CN

4. Unable to run properly

   • have matlab interactive error, like unable to use matlab script in Labview close both the Labview and matlab. Restart the matlab first, and start labview after matlab under fully running!

     Note: the restart sequence matters!

   • could not detect the face region 🔥 库文件运行异常》运行安装目录下的例程 E:\Labview2017\examples\Vision\NIVision OpenCV Utilities\OpenCV Face Detection.vi

Details

• PCA feature extraction

  1. Transform to gray scale image
  2. resize to [200, 200] & reshape to [200x200, 1] put as columns vectors into matrix T
  3. compute mean column vector in matrix T & each vectors minus mean vectors to get centric matrix A

• rgb2gray

  Change from RGB space to YUV or YCbCr color space and use Y value as gray-scale image.

  rgb2gray 通过计算 R、G 和 B 分量的加权和，将 RGB 值转换为灰度值: $0.2989 * R + 0.5870 * G + 0.1140 * B $

  • YUV space

    mainly consider the simplicity to transform the color space, which usually used be compatible Black&White TV. $$ Y = 0.2989 * R + 0.5870 * G + 0.1140 * B \ U = B - Y \ V = R - Y \ whereYisbrightness,U,Viscolor~difference. $$

  Similarly, Y value in YCbCr Space is also computed by this equation, denotes the brightness.