{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import tensorflow as tf\n",
    "from tensorflow.keras.preprocessing.image import ImageDataGenerator\n",
    "import os\n",
    "from tensorflow.keras import layers, models\n",
    "from tensorflow.keras.optimizers import Adam\n",
    "import numpy as np\n",
    "from sklearn.metrics import f1_score, precision_score, recall_score, confusion_matrix"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Found 32 images belonging to 4 classes.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Found 8 images belonging to 4 classes.\n",
      "Class indices: {'Cyst': 0, 'Normal': 1, 'Stone': 2, 'Tumor': 3}\n",
      "done\n"
     ]
    }
   ],
   "source": [
    "# Define paths for the dataset\n",
    "base_dir = '../images'\n",
    "\n",
    "# Create ImageDataGenerators for training, validation, and testing\n",
    "data_gen = ImageDataGenerator(\n",
    "    rescale=1.0/255, # Normalize pixel values\n",
    "    validation_split=0.2 # Split for validation\n",
    ")\n",
    "\n",
    "# Load training data\n",
    "train_data = data_gen.flow_from_directory(\n",
    "    base_dir,\n",
    "    target_size=(150, 150),\n",
    "    batch_size=2,\n",
    "    class_mode='categorical',\n",
    "    subset='training'\n",
    ")\n",
    "\n",
    "# Load validation data\n",
    "val_data = data_gen.flow_from_directory(\n",
    "    base_dir,\n",
    "    target_size=(150, 150),\n",
    "    batch_size=2,\n",
    "    class_mode='categorical',\n",
    "    subset='validation'\n",
    ")\n",
    "\n",
    "# Print class indices\n",
    "print(\"Class indices:\", train_data.class_indices)\n",
    "print('done')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "input_shape = (150, 150, 3)  # 750x750 RGB images\n",
    "num_classes = 4"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 109,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Create the CNN model\n",
    "model = models.Sequential([\n",
    "    # Input layer\n",
    "    layers.Input(shape=input_shape),\n",
    "    \n",
    "    # First Convolutional Block\n",
    "    layers.Conv2D(32, kernel_size=(3, 3), activation='relu', padding='same'),\n",
    "    layers.BatchNormalization(),\n",
    "    layers.MaxPooling2D(pool_size=(2, 2)),\n",
    "    \n",
    "    # Second Convolutional Block\n",
    "    layers.Conv2D(64, kernel_size=(3, 3), activation='relu', padding='same'),\n",
    "    layers.BatchNormalization(),\n",
    "    layers.MaxPooling2D(pool_size=(2, 2)),\n",
    "    \n",
    "    # Third Convolutional Block\n",
    "    layers.Conv2D(128, kernel_size=(3, 3), activation='relu', padding='same'),\n",
    "    layers.BatchNormalization(),\n",
    "    layers.MaxPooling2D(pool_size=(2, 2)),\n",
    "    \n",
    "    # Fourth Convolutional Block\n",
    "    layers.Conv2D(256, kernel_size=(3, 3), activation='relu', padding='same'),\n",
    "    layers.BatchNormalization(),\n",
    "    layers.MaxPooling2D(pool_size=(2, 2)),\n",
    "    \n",
    "    # Fifth Convolutional Block\n",
    "    layers.Conv2D(512, kernel_size=(3, 3), activation='relu', padding='same'),\n",
    "    layers.BatchNormalization(),\n",
    "    layers.MaxPooling2D(pool_size=(2, 2)),\n",
    "    \n",
    "    # Fully Connected Layers\n",
    "    layers.Flatten(),\n",
    "    # layers.Dense(1024, activation='relu'),  # Adjusted to match the input shape\n",
    "    # layers.Dropout(0.5),\n",
    "    # layers.Dense(128, activation='relu'),\n",
    "    # layers.Dropout(0.5),\n",
    "    # \n",
    "    # Output Layer\n",
    "    layers.Dense(num_classes, activation='softmax')\n",
    "])\n",
    "\n",
    "# Compile the model\n",
    "model.compile(optimizer=Adam(learning_rate=0.00001),\n",
    "              loss='categorical_crossentropy',  # Use 'categorical_crossentropy' for one-hot encoded labels\n",
    "              metrics=['accuracy', 'f1_score'])\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 110,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/30\n",
      "\u001b[1m16/16\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 158ms/step - accuracy: 0.5123 - f1_score: 0.4374 - loss: 1.7693 - val_accuracy: 0.2500 - val_f1_score: 0.1000 - val_loss: 1.3858\n",
      "Epoch 2/30\n",
      "\u001b[1m16/16\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 127ms/step - accuracy: 0.9428 - f1_score: 0.8688 - loss: 0.0977 - val_accuracy: 0.2500 - val_f1_score: 0.1000 - val_loss: 1.3971\n",
      "Epoch 3/30\n",
      "\u001b[1m16/16\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 144ms/step - accuracy: 1.0000 - f1_score: 0.9412 - loss: 0.0092 - val_accuracy: 0.2500 - val_f1_score: 0.1000 - val_loss: 1.4072\n",
      "Epoch 4/30\n",
      "\u001b[1m16/16\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 141ms/step - accuracy: 1.0000 - f1_score: 0.9265 - loss: 0.0102 - val_accuracy: 0.2500 - val_f1_score: 0.1000 - val_loss: 1.4199\n",
      "Epoch 5/30\n",
      "\u001b[1m16/16\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 140ms/step - accuracy: 1.0000 - f1_score: 0.9559 - loss: 0.0051 - val_accuracy: 0.2500 - val_f1_score: 0.1000 - val_loss: 1.4359\n",
      "Epoch 6/30\n",
      "\u001b[1m16/16\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 140ms/step - accuracy: 1.0000 - f1_score: 0.9118 - loss: 0.0038 - val_accuracy: 0.2500 - val_f1_score: 0.1000 - val_loss: 1.4536\n",
      "Epoch 7/30\n",
      "\u001b[1m16/16\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 130ms/step - accuracy: 1.0000 - f1_score: 0.9265 - loss: 0.0032 - val_accuracy: 0.2500 - val_f1_score: 0.1000 - val_loss: 1.4734\n",
      "Epoch 8/30\n",
      "\u001b[1m16/16\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 136ms/step - accuracy: 1.0000 - f1_score: 0.8824 - loss: 0.0044 - val_accuracy: 0.2500 - val_f1_score: 0.1000 - val_loss: 1.4929\n",
      "Epoch 9/30\n",
      "\u001b[1m16/16\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 138ms/step - accuracy: 1.0000 - f1_score: 0.9265 - loss: 0.0031 - val_accuracy: 0.2500 - val_f1_score: 0.1000 - val_loss: 1.5133\n",
      "Epoch 10/30\n",
      "\u001b[1m16/16\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 129ms/step - accuracy: 1.0000 - f1_score: 0.9706 - loss: 0.0024 - val_accuracy: 0.2500 - val_f1_score: 0.1000 - val_loss: 1.5346\n",
      "Epoch 11/30\n",
      "\u001b[1m16/16\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 127ms/step - accuracy: 1.0000 - f1_score: 0.8971 - loss: 0.0029 - val_accuracy: 0.2500 - val_f1_score: 0.1111 - val_loss: 1.5535\n",
      "Epoch 12/30\n",
      "\u001b[1m16/16\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 135ms/step - accuracy: 1.0000 - f1_score: 0.9412 - loss: 0.0029 - val_accuracy: 0.3750 - val_f1_score: 0.2429 - val_loss: 1.5660\n",
      "Epoch 13/30\n",
      "\u001b[1m16/16\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 141ms/step - accuracy: 1.0000 - f1_score: 0.9118 - loss: 0.0014 - val_accuracy: 0.3750 - val_f1_score: 0.2429 - val_loss: 1.5759\n",
      "Epoch 14/30\n",
      "\u001b[1m16/16\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 129ms/step - accuracy: 1.0000 - f1_score: 0.8824 - loss: 0.0016 - val_accuracy: 0.3750 - val_f1_score: 0.2429 - val_loss: 1.5809\n",
      "Epoch 15/30\n",
      "\u001b[1m16/16\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 140ms/step - accuracy: 1.0000 - f1_score: 0.9412 - loss: 0.0014 - val_accuracy: 0.3750 - val_f1_score: 0.2429 - val_loss: 1.5782\n",
      "Epoch 16/30\n",
      "\u001b[1m16/16\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 132ms/step - accuracy: 1.0000 - f1_score: 0.9559 - loss: 0.0011 - val_accuracy: 0.2500 - val_f1_score: 0.1250 - val_loss: 1.5731\n",
      "Epoch 17/30\n",
      "\u001b[1m16/16\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 133ms/step - accuracy: 1.0000 - f1_score: 0.9559 - loss: 0.0015 - val_accuracy: 0.3750 - val_f1_score: 0.2429 - val_loss: 1.5475\n",
      "Epoch 18/30\n",
      "\u001b[1m16/16\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 130ms/step - accuracy: 1.0000 - f1_score: 0.9412 - loss: 0.0010 - val_accuracy: 0.3750 - val_f1_score: 0.2429 - val_loss: 1.5120\n",
      "Epoch 19/30\n",
      "\u001b[1m16/16\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 137ms/step - accuracy: 1.0000 - f1_score: 0.9559 - loss: 0.0010 - val_accuracy: 0.2500 - val_f1_score: 0.1250 - val_loss: 1.4640\n",
      "Epoch 20/30\n",
      "\u001b[1m16/16\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 142ms/step - accuracy: 1.0000 - f1_score: 0.9559 - loss: 0.0013 - val_accuracy: 0.3750 - val_f1_score: 0.2429 - val_loss: 1.4229\n",
      "Epoch 21/30\n",
      "\u001b[1m16/16\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 135ms/step - accuracy: 1.0000 - f1_score: 0.9118 - loss: 9.3909e-04 - val_accuracy: 0.3750 - val_f1_score: 0.2429 - val_loss: 1.3761\n",
      "Epoch 22/30\n",
      "\u001b[1m16/16\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 133ms/step - accuracy: 1.0000 - f1_score: 0.9118 - loss: 8.4527e-04 - val_accuracy: 0.5000 - val_f1_score: 0.3333 - val_loss: 1.3416\n",
      "Epoch 23/30\n",
      "\u001b[1m16/16\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 136ms/step - accuracy: 1.0000 - f1_score: 0.9559 - loss: 8.5071e-04 - val_accuracy: 0.5000 - val_f1_score: 0.3333 - val_loss: 1.3063\n",
      "Epoch 24/30\n",
      "\u001b[1m16/16\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 134ms/step - accuracy: 1.0000 - f1_score: 0.9559 - loss: 7.2937e-04 - val_accuracy: 0.5000 - val_f1_score: 0.3429 - val_loss: 1.2619\n",
      "Epoch 25/30\n",
      "\u001b[1m16/16\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 135ms/step - accuracy: 1.0000 - f1_score: 0.9265 - loss: 6.2996e-04 - val_accuracy: 0.5000 - val_f1_score: 0.3429 - val_loss: 1.1914\n",
      "Epoch 26/30\n",
      "\u001b[1m16/16\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 135ms/step - accuracy: 1.0000 - f1_score: 0.9265 - loss: 7.7824e-04 - val_accuracy: 0.5000 - val_f1_score: 0.3429 - val_loss: 1.1314\n",
      "Epoch 27/30\n",
      "\u001b[1m16/16\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 135ms/step - accuracy: 1.0000 - f1_score: 0.8971 - loss: 6.3872e-04 - val_accuracy: 0.6250 - val_f1_score: 0.5333 - val_loss: 1.0617\n",
      "Epoch 28/30\n",
      "\u001b[1m16/16\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 134ms/step - accuracy: 1.0000 - f1_score: 0.9412 - loss: 6.7060e-04 - val_accuracy: 0.7500 - val_f1_score: 0.7333 - val_loss: 1.0080\n",
      "Epoch 29/30\n",
      "\u001b[1m16/16\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 140ms/step - accuracy: 1.0000 - f1_score: 0.9559 - loss: 6.5673e-04 - val_accuracy: 0.7500 - val_f1_score: 0.7333 - val_loss: 0.9536\n",
      "Epoch 30/30\n",
      "\u001b[1m16/16\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 135ms/step - accuracy: 1.0000 - f1_score: 0.9412 - loss: 7.8768e-04 - val_accuracy: 0.7500 - val_f1_score: 0.7333 - val_loss: 0.9146\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:You are saving your model as an HDF5 file via `model.save()` or `keras.saving.save_model(model)`. This file format is considered legacy. We recommend using instead the native Keras format, e.g. `model.save('my_model.keras')` or `keras.saving.save_model(model, 'my_model.keras')`. \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model saved as medical_classifier.h5\n"
     ]
    }
   ],
   "source": [
    "# Train the model\n",
    "history = model.fit(\n",
    "    train_data,\n",
    "    steps_per_epoch=len(train_data),\n",
    "    epochs=30,\n",
    "    validation_data=val_data,\n",
    "    validation_steps=len(val_data)\n",
    ")\n",
    "\n",
    "# Save the model\n",
    "model.save('medical_classifier.h5')\n",
    "print(\"Model saved as medical_classifier.h5\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\u001b[1m4/4\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 32ms/step\n",
      "F1 Score on validation data: 0.35\n"
     ]
    }
   ],
   "source": [
    "\n",
    "val_data.reset()\n",
    "predictions = model.predict(val_data, steps=len(val_data), verbose=1)\n",
    "y_pred = np.argmax(predictions, axis=1)\n",
    "y_true = val_data.classes\n",
    "f1 = f1_score(y_true, y_pred, average='weighted')\n",
    "print(\"F1 Score on validation data:\", f1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 111,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\u001b[1m4/4\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 25ms/step\n",
      "F1 Score on validation data: 0.3666666666666667\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import f1_score\n",
    "import numpy as np\n",
    "from tensorflow.keras.preprocessing import image\n",
    "\n",
    "# Calculate F1 score on validation data\n",
    "val_data.reset()\n",
    "predictions = model.predict(val_data, steps=len(val_data), verbose=1)\n",
    "y_pred = np.argmax(predictions, axis=1)\n",
    "y_true = val_data.classes\n",
    "f1 = f1_score(y_true, y_pred, average='weighted')\n",
    "print(\"F1 Score on validation data:\", f1)\n",
    "\n",
    "# Test the model on a random image\n",
    "def test_random_image(img_path):\n",
    "    img = image.load_img(img_path, target_size=(150, 150))\n",
    "    img_array = image.img_to_array(img)\n",
    "    img_array = np.expand_dims(img_array, axis=0)\n",
    "    img_array /= 255.0\n",
    "\n",
    "    prediction = model.predict(img_array)\n",
    "    predicted_class = np.argmax(prediction, axis=1)\n",
    "    class_indices = {v: k for k, v in train_data.class_indices.items()}\n",
    "    predicted_label = class_indices[predicted_class[0]]\n",
    "\n",
    "    print(f\"Predicted class: {predicted_label}\")\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 114,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\u001b[1m4/4\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 29ms/step\n",
      "F1 Score on validation data: 0.3666666666666667\n",
      "\u001b[1m1/1\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 48ms/step\n",
      "Predicted class: Cyst\n"
     ]
    }
   ],
   "source": [
    "val_data.reset()\n",
    "predictions = model.predict(val_data, steps=len(val_data), verbose=1)\n",
    "y_pred = np.argmax(predictions, axis=1)\n",
    "y_true = val_data.classes\n",
    "f1 = f1_score(y_true, y_pred, average='weighted')\n",
    "print(\"F1 Score on validation data:\", f1)\n",
    "random_image_path = os.path.join(base_dir, 'test', 'Cyst', 'Cyst- (18).jpg')  # Replace 'class_name' and 'random_image.jpg' with actual values\n",
    "test_random_image(random_image_path)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 118,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\u001b[1m1/1\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 50ms/step\n",
      "Predicted class: Normal\n"
     ]
    }
   ],
   "source": [
    "random_image_path = os.path.join(base_dir, 'test', 'Normal', 'Normal- (286).jpg')  # Replace 'class_name' and 'random_image.jpg' with actual values\n",
    "test_random_image(random_image_path)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 120,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\u001b[1m1/1\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 51ms/step\n",
      "Predicted class: Stone\n"
     ]
    }
   ],
   "source": [
    "random_image_path = os.path.join(base_dir, 'test', 'Stone', 'Stone- (62).jpg')  # Replace 'class_name' and 'random_image.jpg' with actual values\n",
    "test_random_image(random_image_path)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 122,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\u001b[1m1/1\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 56ms/step\n",
      "Predicted class: Tumor\n"
     ]
    }
   ],
   "source": [
    "random_image_path = os.path.join(base_dir, 'test', 'Tumor', 'Tumor- (54).jpg')  # Replace 'class_name' and 'random_image.jpg' with actual values\n",
    "test_random_image(random_image_path)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 132,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Compiled the loaded model, but the compiled metrics have yet to be built. `model.compile_metrics` will be empty until you train or evaluate the model.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\u001b[1m1/1\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 55ms/step\n",
      "Predicted class: Tumor\n"
     ]
    }
   ],
   "source": [
    "r_img_path = os.path.join(base_dir, 'test', 'Tumor', 'Tumor- (44).jpg')  # Replace 'class_name' and 'random_image.jpg' with actual values\n",
    "import_model = tf.keras.models.load_model('./medical_classifier.h5')\n",
    "img = image.load_img(r_img_path, target_size=(150, 150))\n",
    "img_array = image.img_to_array(img)\n",
    "img_array = np.expand_dims(img_array, axis=0)\n",
    "img_array /= 255.0\n",
    "\n",
    "prediction = model.predict(img_array)\n",
    "predicted_class = np.argmax(prediction, axis=1)\n",
    "class_indices = {v: k for k, v in train_data.class_indices.items()}\n",
    "predicted_label = class_indices[predicted_class[0]]\n",
    "\n",
    "print(f\"Predicted class: {predicted_label}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "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": "3.10.11"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}