detector_testing.ipynb

{
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "FMrH0ZvUYoum"
      },
      "source": [
        "# PIXELATED IMAGE DETECTION MODEL TESTING"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "LqNGnqHXYour"
      },
      "source": [
        "## UPGRADING TO LATEST LIBRARIES\n",
        "\n",
        "For our model to work, we need to upgrade this libraries"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "19RbBjwWg4NO"
      },
      "outputs": [],
      "source": [
        "%pip install --upgrade pip\n",
        "%pip install --upgrade tensorflow\n",
        "%pip install --upgrade scikit-learn\n",
        "%pip install --upgrade pydot\n",
        "%pip install --upgrade matplotlib\n",
        "\n",
        "# # IN CASE YOU ARE FACING ISSUES WHILE LOADING OR IMPORTING LIBRARIES YOU CAN\n",
        "# # USE THE VERSIONS OF THESE LIBRARIES ON WHICH OUR MODEL WAS INITIALLY TRAINED\n",
        "\n",
        "# %pip install --upgrade pip==24.1.1\n",
        "# %pip install --upgrade ipykernel==5.5.6\n",
        "# %pip install --upgrade numpy==1.25.2\n",
        "# %pip install --upgrade pandas==2.0.3\n",
        "# %pip install --upgrade tensorflow==2.16.2\n",
        "# %pip install --upgrade keras==3.4.1\n",
        "# %pip install --upgrade scikit-learn==1.5.0\n",
        "# %pip install --upgrade matplotlib==3.9.0\n",
        "# %pip install --upgrade pydot==2.0.0"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "947PiVx2Youx"
      },
      "source": [
        "## LOADING THE MODEL"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "IElOro0_gRhU"
      },
      "outputs": [],
      "source": [
        "import os\n",
        "import numpy as np\n",
        "import tensorflow as tf\n",
        "import tensorflow.keras as tfk\n",
        "from sklearn.metrics import ConfusionMatrixDisplay, accuracy_score, precision_score, recall_score, f1_score"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "t72mWdwXgRhX"
      },
      "outputs": [],
      "source": [
        "detector = tfk.models.load_model('./detector_005_1_50.keras')"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "sI3jBduCYou1"
      },
      "source": [
        "## LOADING THE DATASET TO FEED TO MODEL"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "iw70onjdgRhY"
      },
      "outputs": [],
      "source": [
        "BATCH_SIZE = 1      # use  1 if all the test images don't have same size\n",
        "                    # use >1 if all the test images are of same size\n",
        "\n",
        "\n",
        "NON_PIXELATED = 0\n",
        "PIXELATED = 1\n",
        "class_names = ['Non-Pixelated', 'Pixelated']"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "OQD9tWMBYou3"
      },
      "source": [
        "### TODO: IMAGE PATHS (IMPORTANT)\n",
        "\n",
        "For pixelated image detction model testing we require two lists:\n",
        "- `image_paths` : this contains all the image paths eg. ['./path/to/image1.jpg', './path/to/image2.png', ...]\n",
        "- `image_labels`: this containts the labels eg. [1, 0, 0, ...]\n",
        "\n",
        "So please implement the following cell according to how your dataset is organized"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "9K1KdTOQgRhY"
      },
      "outputs": [],
      "source": [
        "# image_dir = './un_mod_ultimate/'\n",
        "# image_basenames = sorted(os.listdir(image_dir))\n",
        "\n",
        "# image_paths = [os.path.join(image_dir, bname) for bname in image_basenames]\n",
        "# image_labels = [NON_PIXELATED if bname.split('_')[1][0] == 'r' else PIXELATED for bname in image_basenames]"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "m0pacgHHYou6",
        "outputId": "5ee3fad3-778c-4a3b-84fa-ae0ca40e5bba"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "1028\n",
            "1028\n",
            "['./un_mod_ultimate/800001019_d3.png', './un_mod_ultimate/800001020_d8.png', './un_mod_ultimate/800001021_d2.png', './un_mod_ultimate/800001022_real.png', './un_mod_ultimate/800001023_real.png', './un_mod_ultimate/800001024_d6.png', './un_mod_ultimate/800001025_d3.png', './un_mod_ultimate/800001026_j10.png', './un_mod_ultimate/800001027_d6.png', './un_mod_ultimate/800001028_d2.png']\n",
            "[1, 1, 1, 0, 0, 1, 1, 1, 1, 1]\n"
          ]
        }
      ],
      "source": [
        "# check if your implementation working correctly\n",
        "\n",
        "print(len(image_paths))\n",
        "print(len(image_labels))\n",
        "\n",
        "print(image_paths[-10:])\n",
        "print(image_labels[-10:])"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "UqgcHvvrYou7"
      },
      "source": [
        "### DATASET OBJECT CREATION\n",
        "\n",
        "this function loads the image from corresponding image_path and preprocesses it to have:\n",
        "- pixel values in [0,1], float32 type, 3 channels\n",
        "- pad to have height, width as multiple of 64 (this is required due to model architecture)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "5uqCyjG-gRhY"
      },
      "outputs": [],
      "source": [
        "@tf.function\n",
        "def load_process_image(image_path):\n",
        "    img = tf.io.read_file(image_path)\n",
        "    img = tf.image.decode_image(img, channels=3, expand_animations=False, dtype=tf.float32)\n",
        "\n",
        "    shape = tf.shape(img)\n",
        "    height, width = shape[0], shape[1]\n",
        "    padding_width  = ((width  + 63) // 64) * 64 - width\n",
        "    padding_height = ((height + 63) // 64) * 64 - height\n",
        "    left  = padding_width  // 2\n",
        "    upper = padding_height // 2\n",
        "\n",
        "    return tf.image.pad_to_bounding_box(img, upper, left, height + padding_height, width + padding_width)\n",
        "\n",
        "# create the dataset, batch, prefetch object\n",
        "\n",
        "custom_test_dataset = tf.data.Dataset.from_tensor_slices(image_paths).map(load_process_image, num_parallel_calls=tf.data.AUTOTUNE, deterministic=True)\\\n",
        "                                                                     .batch(batch_size=BATCH_SIZE, num_parallel_calls=tf.data.AUTOTUNE, deterministic=True)\\\n",
        "                                                                     .prefetch(buffer_size=tf.data.AUTOTUNE)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "AxiE2l84You9"
      },
      "source": [
        "## MODEL EVALUATION"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "R6el6bBUYou9",
        "outputId": "d400ac5a-550b-412c-db25-08dd99bfcb5c"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "\u001b[1m257/257\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m114s\u001b[0m 374ms/step\n"
          ]
        }
      ],
      "source": [
        "predictions = detector.predict(custom_test_dataset)\n",
        "pred_labels = tf.where(predictions < 0, NON_PIXELATED, PIXELATED).numpy().flatten()\n",
        "true_labels = np.array(image_labels)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 593
        },
        "id": "CuS8MPYsgRha",
        "outputId": "e252188a-ca82-42bd-8015-ce131a463982"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "\n",
            "Accuracy: 0.8716\n",
            "\n",
            "Class Names    Precision  Recall  F1-Score\n",
            "--------------------------------------------\n",
            "Non-Pixelated    0.8096   0.8781   0.8425\n",
            "Pixelated        0.9172   0.8674   0.8916\n"
          ]
        },
        {
          "data": {
            "image/png": "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",
            "text/plain": [
              "<Figure size 640x480 with 2 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        },
        {
          "data": {
            "image/png": "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",
            "text/plain": [
              "<Figure size 640x480 with 2 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        },
        {
          "data": {
            "image/png": "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",
            "text/plain": [
              "<Figure size 640x480 with 2 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        }
      ],
      "source": [
        "accuracy = accuracy_score(true_labels, pred_labels)\n",
        "precision = precision_score(true_labels, pred_labels, average=None, zero_division=0.0)\n",
        "recall = recall_score(true_labels, pred_labels, average=None, zero_division=0.0)\n",
        "f1 = f1_score(true_labels, pred_labels, average=None)\n",
        "\n",
        "print(f\"\\nAccuracy: {accuracy:.4f}\")\n",
        "print(\"\\nClass Names    Precision  Recall  F1-Score\")\n",
        "print(\"--------------------------------------------\")\n",
        "for class_name, p, r, f in zip(class_names, precision, recall, f1):\n",
        "  print(f\"{class_name:15}  {p:.4f}   {r:.4f}   {f:.4f}\")\n",
        "\n",
        "_ = ConfusionMatrixDisplay.from_predictions(true_labels, pred_labels, display_labels=class_names, normalize='true', cmap='Blues')\n",
        "_ = ConfusionMatrixDisplay.from_predictions(true_labels, pred_labels, display_labels=class_names, normalize= None,  cmap='Blues')\n",
        "_ = ConfusionMatrixDisplay.from_predictions(true_labels, pred_labels, display_labels=class_names, normalize='all',  cmap='Blues')"
      ]
    }
  ],
  "metadata": {
    "accelerator": "GPU",
    "colab": {
      "gpuType": "T4",
      "provenance": []
    },
    "kernelspec": {
      "display_name": "Python 3",
      "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": "3.12.4"
    }
  },
  "nbformat": 4,
  "nbformat_minor": 0
}