ollewelin

///Now with fully connected Logistic Regression network Supervised Learning
///************* Parameters and things regarding fully connected network **************
int fully_conn_backprop =0;
const int C_fully_hidd_nodes = 200;
const int C_fully_out_nodes = 10;
int fully_hidd_nodes = C_fully_hidd_nodes;
int fully_out_nodes = C_fully_out_nodes;
int drop_out_percent = 50;/// 50% dropout percent hidden nodes during training
int verification = 0;
float Error_level=0.0f;

ollewelin

///Stacked Autoencoder 2 layer L1 and L2
///TODO: Not yet show the image representation off each L2 feature projected on a L1 layer image representation
int Pause_cam =0;
#define USE_RASPICAM_INPUT //If you want to use raspicam input data
//#define USE_MNIST_DATABASE// Here read the t10k-images-idx3-ubyte file

//Here use real image from raspicam take random part 10x10 pixel and put in to the autoencoder
//test stacking autoencoder 2 layer 4x input same Layer 1 (L1) filter down to the Layer 2 (L2)
//L2 input nodes = 4x L1 nodes

ollewelin

///Stacked Autoencoder 2 layer L1 and L2
///Hit <?> to read help menu.
///Now USE_IND_NOISE switch ON make more realistoc gabor filter like feature on first layer  
///TODO: Not yet show the image representation off each L2 feature projected on a L1 layer image representation
int Pause_cam =0;
#define USE_RASPICAM_INPUT //If you want to use raspicam input data
//#define USE_MNIST_DATABASE// Here read the t10k-images-idx3-ubyte file

//Here use real image from raspicam take random part 10x10 pixel and put in to the autoencoder
//test stacking autoencoder 2 layer 4x input same Layer 1 (L1) filter down to the Layer 2 (L2)

ollewelin

///Now USE_IND_NOISE switch ON make more realistoc gabor filter like feature 

           /// Fix Bugg replace hidden_node[j] and output_node[i] with Bias_level
           //   change_weight_in2hid[j] = (LearningRate/2) * hidden_node[j] * hid_node_delta[j] + Momentum * change_weight_in2hid[j];
           //   change_weight_hid2out[i] = (LearningRate/2) * output_node[i] * delta_pixel + Momentum * change_weight_hid2out[i];
///#define USE_LIM_BIAS//

//Here use real image from raspicam take random part 10x10 pixel and put in to the autoencoder

const int cam_h = 240;

ollewelin

///Now also Add bias nodes showed in the last 2 patches
const float Bias_level = 1.0f;
const float Bias_w_n_range = -1.0f;
const float Bias_w_p_range = 1.0f;
const float change_bias_weight_range = 0.2f;


//Add visualize hidden nodes
//Add so you can save weight by press <S>
//Autoencoder learning test with raspicam

ollewelin

//Add visualize hidden nodes
//Add so you can save weight by press <S>
//Autoencoder learning test with raspicam
//press <Y> at start It's tested with dataset from
/// t10k-images-idx3-ubyte
/// http://yann.lecun.com/exdb/mnist/

//If you comment out USE_MNIST_DATABASE switch then simpler pattern used from a picture https://blog.webkid.io/datasets-for-machine-learning/
//0..9 28x28 pixel digits pattern mnist.png filname 289x289 orginal
//orginal image taken from

ollewelin

//Test2 improved and also show (visualize) the noise input stimulu 
//Autoencoder learning test with raspicam
//press <Y> at start It's tested with 0..9 28x28 pixel digits pattern mnist.png filname 289x289 orginal
//orginal image taken from
//       https://blog.webkid.io/datasets-for-machine-learning/

//#define USE_RASPICAM_INPUT //If you want to use raspicam input data


#include <opencv2/highgui/highgui.hpp>  // OpenCV window I/O

ollewelin

//Autoencoder learning test with raspicam
//press <Y> at start It's tested with 0..9 28x28 pixel digits pattern mnist.png filname 289x289 orginal
//orginal image taken from
//       https://blog.webkid.io/datasets-for-machine-learning/

//#define USE_RASPICAM_INPUT //If you want to use raspicam input data


#include <opencv2/highgui/highgui.hpp>  // OpenCV window I/O
#include <opencv2/imgproc/imgproc.hpp> // Gaussian Blur

ollewelin

//This code take some posXXX.jpg and verXXX.jpg files in program root dir and tranform images to a folder with name \positive_data\posXXX.jpg verXXX.jpg

//#include <stdio.h>
//#include <unistd.h>
//#include <ctime>
//#include <iostream>
//#include <raspicam/raspicam_cv.h>

#include <opencv2/highgui/highgui.hpp>  // OpenCV window I/O
#include <opencv2/imgproc/imgproc.hpp> // Gaussian Blur

ollewelin

//2017-05-11 fix bugg index was wrong before m32_conv0 = convolute_mat2(&Feature0Kernel[31][0], FE0KSIZESQR, FE0KSIZESQR, m1_0_padded, int (m1_0_padded.cols));// Make a convolution of the image
//2017-05-05 32x32x32 feature stright connection
//2017-05-05 FIX so kernel update weight direct after each pixel step (before there was a sum up all steps togheter dont work prorper). So now the kerenle patches traingin much much faster and better
//the kernels feature now adapt and look's much more like it correspond to the traingin images.
//Add dropout on fully connected HiddenNodes prevent overtraning
//2017-01-17 fix bugg in make_SumChangeFeature0Weights replace c_m1 with m1_conv0
//Ask of loading any reruned training turn
//Rerun fully connected weights (10 times) and lock kernel layer 0 after (2) rerun lock kernel layer 1 after (4) reruns
//Show all kernels in 3 windows
//In this example 6 output nodes and explaned training images set in start of program.