SVM_wood_classification.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "415f3419-6e9b-448c-8376-044799f8bc78",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Accuracy: 0.83\n",
      "Validation Accuracy: 0.82\n",
      "Test Accuracy: 0.82\n",
      "Classification Report:\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "     defects       0.82      1.00      0.90      1194\n",
      "      normal       1.00      0.00      0.00       265\n",
      "\n",
      "    accuracy                           0.82      1459\n",
      "   macro avg       0.91      0.50      0.45      1459\n",
      "weighted avg       0.85      0.82      0.74      1459\n",
      "\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "#Importing Libraries:\n",
    "\n",
    "import numpy as np\n",
    "import os\n",
    "import cv2\n",
    "import pandas as pd\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "from sklearn.svm import SVC\n",
    "from sklearn.metrics import accuracy_score, classification_report\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "\n",
    "\n",
    "# Loading Data\n",
    "\n",
    "dire = \"D:/nlcdata\"\n",
    "categories = ['defects', 'normal']\n",
    "\n",
    "images = []\n",
    "labels = []\n",
    "\n",
    "SIZE = 600\n",
    "\n",
    "for category in categories:\n",
    "    path = os.path.join(dire, category)\n",
    "    label = categories.index(category)\n",
    "    \n",
    "    for img in os.listdir(path):\n",
    "        imgpath = os.path.join(path, img)\n",
    "        tree_img = cv2.imread(imgpath, 0)\n",
    "        try:\n",
    "            tree_img = cv2.resize(tree_img, (SIZE, SIZE))\n",
    "            images.append(tree_img)\n",
    "            labels.append(label)\n",
    "        except Exception as e:\n",
    "            pass\n",
    "\n",
    "images = np.array(images)\n",
    "labels = np.array(labels)\n",
    "\n",
    "\n",
    "\n",
    "#Feature Extraction\n",
    "\n",
    "def feature_extractor(dataset):\n",
    "    image_dataset = pd.DataFrame()\n",
    "    for img in dataset:\n",
    "        glcm = cv2.calcHist([img.astype(np.uint8)], [0], None, [256], [0, 256])\n",
    "        contrast = np.sum((np.arange(glcm.shape[0]) - np.arange(glcm.shape[1])) ** 2 * glcm)\n",
    "        dissimilarity = np.sum(np.abs(np.arange(glcm.shape[0]) - np.arange(glcm.shape[1])) * glcm)\n",
    "        homogeneity = np.sum(glcm / (1 + np.abs(np.arange(glcm.shape[0]) - np.arange(glcm.shape[1]))))\n",
    "        energy = np.sum(glcm ** 2)\n",
    "        \n",
    "        df = pd.DataFrame({\n",
    "            'Energy': energy,\n",
    "            'Diss_sim': dissimilarity,\n",
    "            'Homogen': homogeneity,\n",
    "            'Contrast': contrast\n",
    "        }, index=[0])\n",
    "        \n",
    "        image_dataset = pd.concat([image_dataset, df], ignore_index=True)\n",
    "        \n",
    "    return image_dataset\n",
    "\n",
    "X_for_ML = feature_extractor(images)\n",
    "\n",
    "\n",
    "# Data Splitiing\n",
    "\n",
    "# Train-test split\n",
    "xtrain, xtest, ytrain, ytest = train_test_split(X_for_ML, labels, test_size=0.15, random_state=42)\n",
    "\n",
    "# Further split the data into training and validation sets\n",
    "xtrain, xval, ytrain, yval = train_test_split(xtrain, ytrain, test_size=0.10, random_state=42)\n",
    "\n",
    "\n",
    "\n",
    "# Feature scaling\n",
    "scaler = StandardScaler()\n",
    "xtrain_scaled = scaler.fit_transform(xtrain)\n",
    "xtest_scaled = scaler.transform(xtest)\n",
    "xval_scaled = scaler.transform(xval)  # Scale validation data using the same scaler\n",
    "\n",
    "\n",
    "\n",
    "# Model Initialization and training\n",
    "# Initialize SVM classifier\n",
    "svm_classifier = SVC(kernel='linear', C=1.0)\n",
    "\n",
    "# Train the model\n",
    "svm_classifier.fit(xtrain_scaled, ytrain)\n",
    "\n",
    "# Predictions on validation data\n",
    "y_pred_val = svm_classifier.predict(xval_scaled)\n",
    "\n",
    "# Predictions on test data\n",
    "y_pred_test = svm_classifier.predict(xtest_scaled)\n",
    "\n",
    "\n",
    "\n",
    "#Model evaluation\n",
    "# Evaluate accuracy\n",
    "train_accuracy = accuracy_score(ytrain, svm_classifier.predict(xtrain_scaled))\n",
    "val_accuracy = accuracy_score(yval, y_pred_val)\n",
    "test_accuracy = accuracy_score(ytest, y_pred_test)\n",
    "\n",
    "print(f\"Training Accuracy: {train_accuracy:.2f}\")\n",
    "print(f\"Validation Accuracy: {val_accuracy:.2f}\")\n",
    "print(f\"Test Accuracy: {test_accuracy:.2f}\")\n",
    "\n",
    "# Classification report\n",
    "print(\"Classification Report:\")\n",
    "print(classification_report(ytest, y_pred_test, target_names=categories, zero_division=1))\n",
    "\n",
    "\n",
    "#Visulaization\n",
    "\n",
    "# Plotting decision boundaries for the validation set\n",
    "plt.scatter(xval_scaled[:, 0], xval_scaled[:, 1], c=y_pred_val, cmap='viridis')\n",
    "plt.xlabel('Feature 1')\n",
    "plt.ylabel('Feature 2')\n",
    "plt.title('SVM Decision Boundary (Validation Set)')\n",
    "plt.show()\n",
    "\n",
    "# Plotting decision boundaries for the test set\n",
    "plt.scatter(xtest_scaled[:, 0], xtest_scaled[:, 1], c=y_pred_test, cmap='viridis')\n",
    "plt.xlabel('Feature 1')\n",
    "plt.ylabel('Feature 2')\n",
    "plt.title('SVM Decision Boundary (Test Set)')\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "06d8fc22-d265-43af-bbbc-d5c1465f0484",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.11.8"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}