Project01 Differentiate between Image of Plus n Cross.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Problem:\n",
    "1. Consider Image of size 3*3 for plus and cross\n",
    "2. In terms of input plus = [0,1,0,1,1,1,0,1,0], cross = [1,0,1,0,1,0,1,0,1]\n",
    "3. 9 Input, 4 Hidden Neuron and 2 Output Neuron\n",
    "\n",
    "\n",
    "# Algorithm:\n",
    "1. Error at Output Layer = (Out_exp - Oct_pre) * Oct_pre * (1 - Oct_pre)\n",
    "2. Re calculate Weight = Weight + learning * error* input\n",
    "3. Error at Hidden Layer = w_11 * error1 + w_12 * error2 \n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Giving input as cross\n",
      "0.11537446108962761 0.8877307596037532\n",
      "Giving input as plus\n",
      "0.8777119697824838 0.1199768847995574\n"
     ]
    }
   ],
   "source": [
    "import numpy as np\n",
    "import os\n",
    "import random\n",
    "import cv2, os\n",
    "import matplotlib.pyplot as plt\n",
    "import math\n",
    "\n",
    "lr = 1 #learning rate\n",
    "#Initialize Bias and Weights\n",
    "bias = [random.random() for x in range(4)]\n",
    "bias_out = [random.random() for x in range(2)]\n",
    "h, w = 10, 5\n",
    "weight_hid = [[random.random() for x in range(w)] for y in range(h)] \n",
    "h, w = 5, 3\n",
    "weight_out = [[random.random() for x in range(w)] for y in range(h)] \n",
    "\n",
    "def flatten_image(image_path):\n",
    "    image = cv2.imread(image_path)\n",
    "    image_grayscale = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY )\n",
    "    size=(3,3)\n",
    "    resized_image = cv2.resize(image_grayscale, size)\n",
    "    image_flatten = resized_image.flatten()\n",
    "    for i in range(len(image_flatten)):\n",
    "        if image_flatten[i] > 150:\n",
    "            image_flatten[i] = 1\n",
    "        else:\n",
    "            image_flatten[i] = 0\n",
    "    return image_flatten     \n",
    "\n",
    "def sigmoid(x):\n",
    "    return 1/(1+math.exp(-x))\n",
    "\n",
    "def hidden_layer(x):\n",
    "    hid_outputP = [0,0,0,0]\n",
    "    #hid_error = [0,0,0,0]\n",
    "\n",
    "    for i in range(0,4):\n",
    "        hid_outputP[i] = bias[i]*weight_hid[0][i+1]\n",
    "        for j in range(0,9):\n",
    "            hid_outputP[i] =  hid_outputP[i] + x[j]*weight_hid[j+1][i+1]\n",
    "          \n",
    "        hid_outputP[i] = sigmoid(hid_outputP[i])\n",
    "    return hid_outputP\n",
    "\n",
    "def output_layer(h, result, plus):\n",
    "    outputP = [0,0]\n",
    "    error = [0,0]\n",
    "    for i in range(0,2):\n",
    "        outputP[i] = bias_out[i]\n",
    "        for j in range(0,4):\n",
    "            outputP[i] =  outputP[i] + h[j]*weight_out[j+1][i+1]\n",
    "\n",
    "    outputP[0] = sigmoid(outputP[0])\n",
    "    outputP[1] = sigmoid(outputP[1])\n",
    "    #Calculate Error at Output Layer\n",
    "    error[0] = (result[0] - outputP[0])*outputP[0]*(1 - outputP[0])\n",
    "    error[1] = (result[1] - outputP[1])*outputP[1]*(1 - outputP[1])\n",
    "\n",
    "    #Calculate Error at Hidden Layer\n",
    "    for i in range(1,5):\n",
    "        error_hid[i] = (error[0]*weight_out[i][1] + error[1]*weight_out[i][2])*h[i-1]*(1 - h[i-1])\n",
    "    #Recalculate Weight to Output Layer\n",
    "    for i in range(1,5):\n",
    "        weight_out[i][1] =  weight_out[i][1] + error[0]*h[i-1]*lr\n",
    "        weight_out[i][2] =  weight_out[i][2] + error[1]*h[i-1]*lr\n",
    "        \n",
    "    #Recalculate Weight to Hidden Layer\n",
    "    for i in range(1,5):\n",
    "        for j in range(1,10):\n",
    "            weight_hid[j][i] = weight_hid[j][i] + error_hid[i]*plus[j-1]\n",
    "\n",
    "def predict(input):\n",
    "    input_hid = hidden_layer(input)\n",
    "    outputP = [0,0]\n",
    "    for i in range(0,2):\n",
    "        outputP[i] = bias_out[i]\n",
    "        for j in range(0,4):\n",
    "            outputP[i] =  outputP[i] + input_hid[j]*weight_out[j+1][i+1]\n",
    "\n",
    "    outputP[0] = sigmoid(outputP[0])\n",
    "    outputP[1] = sigmoid(outputP[1])    \n",
    "    print(outputP[0],outputP[1])  \n",
    "    \n",
    "#plus = [0,1,0,1,1,1,0,1,0]\n",
    "result_plus = [0.9, 0.1]\n",
    "#cross = [1,0,1,0,1,0,1,0,1]\n",
    "result_cross = [0.1, 0.9]\n",
    "error_hid = [0,0,0,0,0]\n",
    "plus_path = 'Data/NN/plus.jpg'\n",
    "plus = flatten_image(plus_path)\n",
    "plus_cross = 'Data/NN/cross.jpg'\n",
    "cross = flatten_image(plus_cross)\n",
    "\n",
    "for i in range(200) :\n",
    "    #print(\"Running loop i=%s\"%i)    \n",
    "    H = hidden_layer(plus) \n",
    "    output_layer(H, result_plus, plus)\n",
    "    H = hidden_layer(cross) \n",
    "    output_layer(H, result_cross,cross)\n",
    "    \n",
    "print(\"Giving input as cross\")\n",
    "predict(cross)\n",
    "print(\"Giving input as plus\")\n",
    "predict(plus)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[{'weights': [0.6394267984578837, 0.025010755222666936, 0.27502931836911926, 0.22321073814882275, 0.7364712141640124, 0.6766994874229113, 0.8921795677048454, 0.08693883262941615, 0.4219218196852704, 0.029797219438070344], 'output': 0.8554655884280492}, {'weights': [0.21863797480360336, 0.5053552881033624, 0.026535969683863625, 0.1988376506866485, 0.6498844377795232, 0.5449414806032167, 0.2204406220406967, 0.5892656838759087, 0.8094304566778266, 0.006498759678061017], 'output': 0.9237752998328821}, {'weights': [0.8058192518328079, 0.6981393949882269, 0.3402505165179919, 0.15547949981178155, 0.9572130722067812, 0.33659454511262676, 0.09274584338014791, 0.09671637683346401, 0.8474943663474598, 0.6037260313668911], 'output': 0.94520841996855}, {'weights': [0.8071282732743802, 0.7297317866938179, 0.5362280914547007, 0.9731157639793706, 0.3785343772083535, 0.552040631273227, 0.8294046642529949, 0.6185197523642461, 0.8617069003107772, 0.577352145256762], 'output': 0.9787369992681083}]\n",
      "[{'weights': [0.7045718362149235, 0.045824383655662215, 0.22789827565154686, 0.28938796360210717, 0.0797919769236275], 'output': 0.7726800302787462}, {'weights': [0.23279088636103018, 0.10100142940972912, 0.2779736031100921, 0.6356844442644002, 0.36483217897008424], 'output': 0.8237829009025937}]\n"
     ]
    }
   ],
   "source": [
    "from random import random\n",
    "from random import seed\n",
    "import math\n",
    "\n",
    "def sigmoid(x):\n",
    "    return 1/(1+math.exp(-x))\n",
    "\n",
    "def init_neuron(input, hidden, output):\n",
    "    neuron = list()\n",
    "    weight_hidden = [{'weights':[ random() for y in range(input + 1)]} for x in range(hidden)]\n",
    "    neuron.append(weight_hidden)\n",
    "    weight_out = [{'weights':[ random() for y in range(hidden + 1)]} for x in range(output)]\n",
    "    neuron.append(weight_out)\n",
    "    return neuron\n",
    "\n",
    "def neuron_summ(weights, input):\n",
    "    summ = weights[-1]\n",
    "    for i in range(len(weights) -1):\n",
    "        summ += weights[i]*input[i]\n",
    "    return summ\n",
    "\n",
    "def fwd_prop(network,x):\n",
    "    \n",
    "    for layer in network:\n",
    "        x_next = []\n",
    "        for neuron in layer:\n",
    "            summ = neuron_summ(neuron['weights'], x)\n",
    "            neuron['output'] = sigmoid(summ)\n",
    "            x_next.append(neuron['output'])\n",
    "        x = x_next\n",
    "\n",
    "seed(42)\n",
    "network = init_neuron(9,4,2)\n",
    "plus = [0,1,0,1,1,1,0,1,0]\n",
    "fwd_prop(network, plus)\n",
    "\n",
    "for layer in network:\n",
    "    print(layer)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}