Introduction - If you have any usage issues, please Google them yourself
Boy wrote on Neural PCA image compression and decompression of the source image after training, can make use of the characteristics of images (eigenvalue and eigenvetex) to other copies of illustrated compression, a very interesting way, and then set when threashold attention values not too large, because it involves the calculation of inverse matrex.
Packet : 113172203pca.rar filelist
pic\DOG64.BMP
pic\KIDS64.BMP
pic\MAN64.BMP
pic\OCEAN64.BMP
pic
Project1.~bpr
Project1.bpr
Project1.cpp
Project1.exe
Project1.obj
Project1.res
Project1.tds
Unit1.~cpp
Unit1.~ddp
Unit1.~dfm
Unit1.~h
Unit1.cpp
Unit1.ddp
Unit1.dfm
Unit1.h
Unit1.obj