├── CNN Report.docx ├── README.md └── CNN_project.ipynb /CNN Report.docx: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/Rahat-karim/CNN/HEAD/CNN Report.docx -------------------------------------------------------------------------------- /README.md: -------------------------------------------------------------------------------- 1 | # CNN 2 | This repo has a project which is based on CNN(convensional neural network) of deep learning in which handwriten numeric numbers were detected. 3 | -------------------------------------------------------------------------------- /CNN_project.ipynb: -------------------------------------------------------------------------------- 1 | { 2 | "nbformat": 4, 3 | "nbformat_minor": 0, 4 | "metadata": { 5 | "colab": { 6 | "provenance": [] 7 | }, 8 | "kernelspec": { 9 | "name": "python3", 10 | "display_name": "Python 3" 11 | }, 12 | "language_info": { 13 | "name": "python" 14 | } 15 | }, 16 | "cells": [ 17 | { 18 | "cell_type": "code", 19 | "execution_count": 1, 20 | "metadata": { 21 | "colab": { 22 | "base_uri": "https://localhost:8080/", 23 | "height": 1000 24 | }, 25 | "id": "0x-ZUS6mJWNG", 26 | "outputId": "2d56b665-d50c-41db-c0ed-9cab1f0ceadd" 27 | }, 28 | "outputs": [ 29 | { 30 | "output_type": "stream", 31 | "name": "stdout", 32 | "text": [ 33 | "Downloading train_32x32.mat...\n", 34 | "train_32x32.mat downloaded.\n", 35 | "Downloading test_32x32.mat...\n", 36 | "test_32x32.mat downloaded.\n", 37 | "Train images shape: (32, 32, 3, 73257)\n", 38 | "Train labels shape: (73257, 1)\n", 39 | "Test images shape: (32, 32, 3, 26032)\n", 40 | "Test labels shape: (26032, 1)\n", 41 | "Reshaped train images: (73257, 32, 32, 3)\n" 42 | ] 43 | }, 44 | { 45 | "output_type": "stream", 46 | "name": "stderr", 47 | "text": [ 48 | "/usr/local/lib/python3.10/dist-packages/keras/src/layers/convolutional/base_conv.py:107: UserWarning: Do not pass an `input_shape`/`input_dim` argument to a layer. When using Sequential models, prefer using an `Input(shape)` object as the first layer in the model instead.\n", 49 | " super().__init__(activity_regularizer=activity_regularizer, **kwargs)\n" 50 | ] 51 | }, 52 | { 53 | "output_type": "stream", 54 | "name": "stdout", 55 | "text": [ 56 | "Epoch 1/10\n", 57 | "\u001b[1m916/916\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m90s\u001b[0m 95ms/step - accuracy: 0.3688 - loss: 1.7989 - val_accuracy: 0.8277 - val_loss: 0.5971\n", 58 | "Epoch 2/10\n", 59 | "\u001b[1m916/916\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m88s\u001b[0m 96ms/step - accuracy: 0.8027 - loss: 0.6669 - val_accuracy: 0.8698 - val_loss: 0.4494\n", 60 | "Epoch 3/10\n", 61 | "\u001b[1m916/916\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m86s\u001b[0m 94ms/step - accuracy: 0.8424 - loss: 0.5359 - val_accuracy: 0.8768 - val_loss: 0.4122\n", 62 | "Epoch 4/10\n", 63 | "\u001b[1m916/916\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m145s\u001b[0m 97ms/step - accuracy: 0.8612 - loss: 0.4734 - val_accuracy: 0.8901 - val_loss: 0.3821\n", 64 | "Epoch 5/10\n", 65 | "\u001b[1m916/916\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m87s\u001b[0m 95ms/step - accuracy: 0.8749 - loss: 0.4278 - val_accuracy: 0.8969 - val_loss: 0.3541\n", 66 | "Epoch 6/10\n", 67 | "\u001b[1m916/916\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m88s\u001b[0m 96ms/step - accuracy: 0.8859 - loss: 0.3840 - val_accuracy: 0.8992 - val_loss: 0.3470\n", 68 | "Epoch 7/10\n", 69 | "\u001b[1m916/916\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m141s\u001b[0m 95ms/step - accuracy: 0.8938 - loss: 0.3584 - val_accuracy: 0.8950 - val_loss: 0.3548\n", 70 | "Epoch 8/10\n", 71 | "\u001b[1m916/916\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m142s\u001b[0m 94ms/step - accuracy: 0.8992 - loss: 0.3411 - val_accuracy: 0.8992 - val_loss: 0.3395\n", 72 | "Epoch 9/10\n", 73 | "\u001b[1m916/916\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m144s\u001b[0m 96ms/step - accuracy: 0.9047 - loss: 0.3270 - val_accuracy: 0.9048 - val_loss: 0.3278\n", 74 | "Epoch 10/10\n", 75 | "\u001b[1m916/916\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m87s\u001b[0m 95ms/step - accuracy: 0.9068 - loss: 0.3167 - val_accuracy: 0.9077 - val_loss: 0.3261\n", 76 | "\u001b[1m814/814\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m15s\u001b[0m 18ms/step - accuracy: 0.8965 - loss: 0.3650\n", 77 | "Test accuracy: 0.9000\n", 78 | "\u001b[1m814/814\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m12s\u001b[0m 15ms/step\n" 79 | ] 80 | }, 81 | { 82 | "output_type": "display_data", 83 | "data": { 84 | "text/plain": [ 85 | "
" 86 | ], 87 | "image/png": "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\n" 88 | }, 89 | "metadata": {} 90 | }, 91 | { 92 | "output_type": "display_data", 93 | "data": { 94 | "text/plain": [ 95 | "
" 96 | ], 97 | "image/png": "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\n" 98 | }, 99 | "metadata": {} 100 | } 101 | ], 102 | "source": [ 103 | "import os\n", 104 | "import numpy as np\n", 105 | "import matplotlib.pyplot as plt\n", 106 | "import tensorflow as tf\n", 107 | "from tensorflow.keras.models import Sequential\n", 108 | "from tensorflow.keras.layers import Conv2D, MaxPooling2D, Flatten, Dense, Dropout\n", 109 | "from sklearn.metrics import confusion_matrix, ConfusionMatrixDisplay\n", 110 | "from tensorflow.keras.utils import to_categorical\n", 111 | "from sklearn.model_selection import train_test_split\n", 112 | "from scipy.io import loadmat\n", 113 | "import requests\n", 114 | "\n", 115 | "# Helper function to download the dataset if not already present\n", 116 | "def download_svhn(filename, url):\n", 117 | " if not os.path.exists(filename):\n", 118 | " print(f\"Downloading {filename}...\")\n", 119 | " response = requests.get(url)\n", 120 | " with open(filename, 'wb') as f:\n", 121 | " f.write(response.content)\n", 122 | " print(f\"{filename} downloaded.\")\n", 123 | "\n", 124 | "# Download the SVHN dataset if not available\n", 125 | "train_url = \"http://ufldl.stanford.edu/housenumbers/train_32x32.mat\"\n", 126 | "test_url = \"http://ufldl.stanford.edu/housenumbers/test_32x32.mat\"\n", 127 | "\n", 128 | "train_file = \"train_32x32.mat\"\n", 129 | "test_file = \"test_32x32.mat\"\n", 130 | "\n", 131 | "download_svhn(train_file, train_url)\n", 132 | "download_svhn(test_file, test_url)\n", 133 | "\n", 134 | "# 1. Load the SVHN dataset\n", 135 | "svhn_train = loadmat(train_file)\n", 136 | "svhn_test = loadmat(test_file)\n", 137 | "\n", 138 | "# Extract images and labels\n", 139 | "X_train = svhn_train['X']\n", 140 | "y_train = svhn_train['y']\n", 141 | "X_test = svhn_test['X']\n", 142 | "y_test = svhn_test['y']\n", 143 | "\n", 144 | "# 2. Check the data shape\n", 145 | "print(\"Train images shape:\", X_train.shape) # (32, 32, 3, num_samples)\n", 146 | "print(\"Train labels shape:\", y_train.shape)\n", 147 | "print(\"Test images shape:\", X_test.shape)\n", 148 | "print(\"Test labels shape:\", y_test.shape)\n", 149 | "\n", 150 | "# Reshape the data to match TensorFlow's expected input\n", 151 | "X_train = np.moveaxis(X_train, -1, 0)\n", 152 | "X_test = np.moveaxis(X_test, -1, 0)\n", 153 | "print(\"Reshaped train images:\", X_train.shape) # Now (num_samples, 32, 32, 3)\n", 154 | "\n", 155 | "# 3. Normalize the data\n", 156 | "X_train = X_train.astype('float32') / 255.0\n", 157 | "X_test = X_test.astype('float32') / 255.0\n", 158 | "\n", 159 | "# Correct labels: In the dataset, the label '10' represents '0', fix it\n", 160 | "y_train[y_train == 10] = 0\n", 161 | "y_test[y_test == 10] = 0\n", 162 | "\n", 163 | "# One-hot encode the labels\n", 164 | "y_train = to_categorical(y_train, 10)\n", 165 | "y_test = to_categorical(y_test, 10)\n", 166 | "\n", 167 | "# 4. Split the train set into training and validation sets\n", 168 | "X_train, X_val, y_train, y_val = train_test_split(X_train, y_train, test_size=0.2, random_state=42)\n", 169 | "\n", 170 | "# 5. Define the CNN model\n", 171 | "model = Sequential([\n", 172 | " Conv2D(32, (3, 3), activation='relu', input_shape=(32, 32, 3)),\n", 173 | " MaxPooling2D(pool_size=(2, 2)),\n", 174 | " Conv2D(64, (3, 3), activation='relu'),\n", 175 | " MaxPooling2D(pool_size=(2, 2)),\n", 176 | " Conv2D(128, (3, 3), activation='relu'),\n", 177 | " MaxPooling2D(pool_size=(2, 2)),\n", 178 | " Flatten(),\n", 179 | " Dense(128, activation='relu'),\n", 180 | " Dropout(0.5),\n", 181 | " Dense(10, activation='softmax')\n", 182 | "])\n", 183 | "\n", 184 | "# Compile the model\n", 185 | "model.compile(optimizer='adam', loss='categorical_crossentropy', metrics=['accuracy'])\n", 186 | "\n", 187 | "# 6. Train the model\n", 188 | "history = model.fit(X_train, y_train, validation_data=(X_val, y_val), epochs=10, batch_size=64)\n", 189 | "\n", 190 | "# 7. Evaluate the model\n", 191 | "test_loss, test_acc = model.evaluate(X_test, y_test)\n", 192 | "print(f\"Test accuracy: {test_acc:.4f}\")\n", 193 | "\n", 194 | "# 8. Generate predictions\n", 195 | "y_pred = model.predict(X_test)\n", 196 | "y_pred_classes = np.argmax(y_pred, axis=1)\n", 197 | "y_true = np.argmax(y_test, axis=1)\n", 198 | "\n", 199 | "# 9. Confusion Matrix\n", 200 | "cm = confusion_matrix(y_true, y_pred_classes)\n", 201 | "disp = ConfusionMatrixDisplay(confusion_matrix=cm)\n", 202 | "disp.plot(cmap='Blues')\n", 203 | "plt.show()\n", 204 | "\n", 205 | "# 10. Visualize test samples with predicted and true labels\n", 206 | "# Choose indices for samples to display\n", 207 | "indices = [0, 1, 2, 3, 4] # You can choose any indices of interest\n", 208 | "\n", 209 | "# Plot the selected test samples\n", 210 | "plt.figure(figsize=(15, 5))\n", 211 | "for i, idx in enumerate(indices):\n", 212 | " plt.subplot(1, len(indices), i + 1)\n", 213 | " plt.imshow(X_test[idx])\n", 214 | " plt.title(f\"True: {y_true[idx]}\\nPred: {y_pred_classes[idx]}\")\n", 215 | " plt.axis('off')\n", 216 | "\n", 217 | "plt.show()" 218 | ] 219 | } 220 | ] 221 | } --------------------------------------------------------------------------------