diff --git a/machine_learning/classification/Adaboost.ipynb b/machine_learning/classification/Adaboost.ipynb
new file mode 100644
index 0000000000000000000000000000000000000000..359d8fab1a81d93015b0e85080445c7d7193b887
--- /dev/null
+++ b/machine_learning/classification/Adaboost.ipynb
@@ -0,0 +1,314 @@
+{
+ "cells": [
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "## Entrainement d'un modèle avec la méthode des SVM"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 1,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "import os\n",
+ "import numpy as np\n",
+ "import random\n",
+ "from PIL import Image"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "### 1) Fonctions de Preprocessing des datasets"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 2,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "AVERAGE_SIZE_IMAGE = (127, 145) # Thanks to the stats, we know that size of bbox will be (127, 145) -> Average size of labels \n",
+ "\n",
+ "def generate_empty_bbox(image_width, image_height):\n",
+ " \"\"\" \n",
+ " Generate an empty box for images without label\n",
+ " \"\"\"\n",
+ " # Generating random coords for the bbox\n",
+ " x_min = random.randint(0, image_width - AVERAGE_SIZE_IMAGE[0])\n",
+ " y_min = random.randint(0, image_height - AVERAGE_SIZE_IMAGE[1])\n",
+ " \n",
+ " # Compute complete coords of the bbox\n",
+ " x_max = x_min + AVERAGE_SIZE_IMAGE[0]\n",
+ " y_max = y_min + AVERAGE_SIZE_IMAGE[1]\n",
+ " \n",
+ " return (x_min, y_min, x_max, y_max)\n",
+ "\n",
+ "def load_data(image_dir, label_dir):\n",
+ " \"\"\" \n",
+ " Create a dict with all the usefull datas of the dataset\n",
+ " datas = {\n",
+ " \"XXXX\" (name of the file) : {\n",
+ " \"img\" : image as an array,\n",
+ " \"labels\" (data of the labels): {\n",
+ " \"X\" index of the label (0,1,...,n) : {\n",
+ " \"name\" : name of the label,\n",
+ " \"coord\" : coord of the label like xmin, ymin, xmax, ymax,\n",
+ " \"img\" : crooped img of the label,\n",
+ " }\n",
+ " }\n",
+ " }\n",
+ " }\n",
+ " \"\"\"\n",
+ " \n",
+ " datas = {}\n",
+ "\n",
+ " for image_file in os.listdir(image_dir):\n",
+ " # Computing name and files paths\n",
+ " image_path = image_dir + '/' + image_file\n",
+ " name = image_file.split('.')[0]\n",
+ " label_path = label_dir + '/' + name + '.csv'\n",
+ " \n",
+ " # Import image as array\n",
+ " image = np.array(Image.open(image_path))\n",
+ "\n",
+ " # Import labels as array \n",
+ " with open(label_path, 'r') as file:\n",
+ " rows = file.readlines()\n",
+ "\n",
+ " label_data = {}\n",
+ " if rows == ['\\n']: # Create a random empty label to balance model\n",
+ " # Create random coords for empty label\n",
+ " xmin, ymin, xmax, ymax = generate_empty_bbox(image.shape[1], image.shape[0])\n",
+ " \n",
+ " # Get the cropped image (as array) of the label\n",
+ " cropped_image = np.array(Image.fromarray(image[ymin:ymax, xmin:xmax]).resize(AVERAGE_SIZE_IMAGE))\n",
+ " \n",
+ " label_data[0] = {\n",
+ " \"name\":\"empty\",\n",
+ " \"coord\": (xmin, ymin, xmax, ymax),\n",
+ " \"img\":cropped_image\n",
+ " }\n",
+ " else:\n",
+ " for i, row in enumerate(rows): # One image can contain several labels\n",
+ " row = row.strip().split(\",\")\n",
+ "\n",
+ " # Compute coords of the label\n",
+ " xmin, ymin, xmax, ymax = map(int, row[0:4])\n",
+ "\n",
+ " # Get the label name\n",
+ " class_name = row[4]\n",
+ "\n",
+ " # Get the cropped image (as array) of the label\n",
+ " cropped_image = np.array(Image.fromarray(image[ymin:ymax, xmin:xmax]).resize(AVERAGE_SIZE_IMAGE))\n",
+ " \n",
+ " # Adding to the json\n",
+ " label_data[i] = {\n",
+ " \"name\":class_name,\n",
+ " \"coord\": (xmin, ymin, xmax, ymax),\n",
+ " \"img\":cropped_image\n",
+ " }\n",
+ "\n",
+ " datas[name] = {\n",
+ " \"img\" : image,\n",
+ " \"labels\" : label_data,\n",
+ " }\n",
+ " \n",
+ " return datas\n",
+ "\n",
+ "# Dict to convert str class name to int\n",
+ "name_to_int = {\n",
+ " \"danger\": 0,\n",
+ " \"interdiction\": 1,\n",
+ " \"obligation\": 2,\n",
+ " \"stop\": 3,\n",
+ " \"ceder\": 4,\n",
+ " \"frouge\": 5,\n",
+ " \"forange\": 6,\n",
+ " \"fvert\": 7,\n",
+ " \"ff\": 8,\n",
+ " \"empty\": 9\n",
+ "}"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "### 2) Fonction de création des datasets"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 3,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "def create_xy(datas):\n",
+ " # Creating arrays with all labels datas & classes\n",
+ " X = []\n",
+ " Y = []\n",
+ "\n",
+ " for name, data in datas.items():\n",
+ " for row in data[\"labels\"].values():\n",
+ " image_as_array = np.array(row[\"img\"]).flatten()\n",
+ " X.append(image_as_array)\n",
+ " Y.append(name_to_int[row[\"name\"]])\n",
+ "\n",
+ " X = np.array(X)\n",
+ " Y = np.array(Y)\n",
+ "\n",
+ " return X, Y"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "### 3) Création des datasets"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 4,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "# Training dataset\n",
+ "datas_train = load_data(\"../../data/train/images\", \"../../data/train/labels\")\n",
+ "X_train, Y_train = create_xy(datas=datas_train)\n",
+ "\n",
+ "# Validation dataset\n",
+ "datas_val = load_data(\"../../data/val/images\", \"../../data/val/labels\")\n",
+ "X_val, Y_val = create_xy(datas=datas_val)"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "### 4) Application de la méthode Adaboost"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 9,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stderr",
+ "output_type": "stream",
+ "text": [
+ "c:\\Users\\victo\\AppData\\Local\\Programs\\Python\\Python312\\Lib\\site-packages\\sklearn\\ensemble\\_weight_boosting.py:519: FutureWarning: The SAMME.R algorithm (the default) is deprecated and will be removed in 1.6. Use the SAMME algorithm to circumvent this warning.\n",
+ " warnings.warn(\n"
+ ]
+ },
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "Taux d'erreur : 0.6302521008403361\n"
+ ]
+ }
+ ],
+ "source": [
+ "from sklearn.ensemble import AdaBoostClassifier\n",
+ "\n",
+ "adaboost_clf = AdaBoostClassifier(n_estimators=10) # To change\n",
+ "adaboost_clf.fit(X_train, Y_train)\n",
+ "y = adaboost_clf.predict(X_val)\n",
+ "\n",
+ "print(f\"Taux d'erreur : {np.mean(y != Y_val)}\")"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "### 5) Test de la méthode Adaboost avec application des caractéristiques HOG"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 6,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "from skimage.feature import hog\n",
+ "from skimage.color import rgb2gray\n",
+ "\n",
+ "def extract_hog(datas):\n",
+ " # Creating X array with all HOG information of images\n",
+ " X = []\n",
+ "\n",
+ " for name, data in datas.items():\n",
+ " for row in data[\"labels\"].values():\n",
+ " image_as_array = np.array(hog(rgb2gray(row[\"img\"]))).flatten()\n",
+ " X.append(image_as_array)\n",
+ "\n",
+ " return np.array(X)\n",
+ "\n",
+ "\n",
+ "# Update training dataset\n",
+ "X_train_HOG = extract_hog(datas=datas_train)\n",
+ "\n",
+ "# Update validation dataset\n",
+ "X_val_HOG = extract_hog(datas=datas_val)"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 7,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stderr",
+ "output_type": "stream",
+ "text": [
+ "c:\\Users\\victo\\AppData\\Local\\Programs\\Python\\Python312\\Lib\\site-packages\\sklearn\\ensemble\\_weight_boosting.py:519: FutureWarning: The SAMME.R algorithm (the default) is deprecated and will be removed in 1.6. Use the SAMME algorithm to circumvent this warning.\n",
+ " warnings.warn(\n"
+ ]
+ },
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "Taux d'erreur : 0.5378151260504201\n"
+ ]
+ }
+ ],
+ "source": [
+ "adaboost_clf = AdaBoostClassifier(n_estimators=10)\n",
+ "adaboost_clf.fit(X_train_HOG, Y_train)\n",
+ "y_HOG = adaboost_clf.predict(X_val_HOG)\n",
+ "\n",
+ "print(f\"Taux d'erreur : {np.mean(y_HOG != Y_val)}\")"
+ ]
+ }
+ ],
+ "metadata": {
+ "kernelspec": {
+ "display_name": "venv",
+ "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": "3.12.2"
+ }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 2
+}
diff --git a/machine_learning/classification/RandomForest.ipynb b/machine_learning/classification/RandomForest.ipynb
new file mode 100644
index 0000000000000000000000000000000000000000..9e9be72cd7f076603c260f7eacdbd5e51e7f6f22
--- /dev/null
+++ b/machine_learning/classification/RandomForest.ipynb
@@ -0,0 +1,348 @@
+{
+ "cells": [
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "## Entrainement d'un modèle avec la méthode RandomTree"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 1,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "import os\n",
+ "import numpy as np\n",
+ "import random\n",
+ "from PIL import Image"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "### 1) Fonctions de Preprocessing des datasets"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 2,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "AVERAGE_SIZE_IMAGE = (127, 145) # Thanks to the stats, we know that size of bbox will be (127, 145) -> Average size of labels \n",
+ "\n",
+ "def generate_empty_bbox(image_width, image_height):\n",
+ " \"\"\" \n",
+ " Generate an empty box for images without label\n",
+ " \"\"\"\n",
+ " # Generating random coords for the bbox\n",
+ " x_min = random.randint(0, image_width - AVERAGE_SIZE_IMAGE[0])\n",
+ " y_min = random.randint(0, image_height - AVERAGE_SIZE_IMAGE[1])\n",
+ " \n",
+ " # Compute complete coords of the bbox\n",
+ " x_max = x_min + AVERAGE_SIZE_IMAGE[0]\n",
+ " y_max = y_min + AVERAGE_SIZE_IMAGE[1]\n",
+ " \n",
+ " return (x_min, y_min, x_max, y_max)\n",
+ "\n",
+ "def load_data(image_dir, label_dir):\n",
+ " \"\"\" \n",
+ " Create a dict with all the usefull datas of the dataset\n",
+ " datas = {\n",
+ " \"XXXX\" (name of the file) : {\n",
+ " \"img\" : image as an array,\n",
+ " \"labels\" (data of the labels): {\n",
+ " \"X\" index of the label (0,1,...,n) : {\n",
+ " \"name\" : name of the label,\n",
+ " \"coord\" : coord of the label like xmin, ymin, xmax, ymax,\n",
+ " \"img\" : crooped img of the label,\n",
+ " }\n",
+ " }\n",
+ " }\n",
+ " }\n",
+ " \"\"\"\n",
+ " \n",
+ " datas = {}\n",
+ "\n",
+ " for image_file in os.listdir(image_dir):\n",
+ " # Computing name and files paths\n",
+ " image_path = image_dir + '/' + image_file\n",
+ " name = image_file.split('.')[0]\n",
+ " label_path = label_dir + '/' + name + '.csv'\n",
+ " \n",
+ " # Import image as array\n",
+ " image = np.array(Image.open(image_path))\n",
+ "\n",
+ " # Import labels as array \n",
+ " with open(label_path, 'r') as file:\n",
+ " rows = file.readlines()\n",
+ "\n",
+ " label_data = {}\n",
+ " if rows == ['\\n']: # Create a random empty label to balance model\n",
+ " # Create random coords for empty label\n",
+ " xmin, ymin, xmax, ymax = generate_empty_bbox(image.shape[1], image.shape[0])\n",
+ " \n",
+ " # Get the cropped image (as array) of the label\n",
+ " cropped_image = np.array(Image.fromarray(image[ymin:ymax, xmin:xmax]).resize(AVERAGE_SIZE_IMAGE))\n",
+ " \n",
+ " label_data[0] = {\n",
+ " \"name\":\"empty\",\n",
+ " \"coord\": (xmin, ymin, xmax, ymax),\n",
+ " \"img\":cropped_image\n",
+ " }\n",
+ " else:\n",
+ " for i, row in enumerate(rows): # One image can contain several labels\n",
+ " row = row.strip().split(\",\")\n",
+ "\n",
+ " # Compute coords of the label\n",
+ " xmin, ymin, xmax, ymax = map(int, row[0:4])\n",
+ "\n",
+ " # Get the label name\n",
+ " class_name = row[4]\n",
+ "\n",
+ " # Get the cropped image (as array) of the label\n",
+ " cropped_image = np.array(Image.fromarray(image[ymin:ymax, xmin:xmax]).resize(AVERAGE_SIZE_IMAGE))\n",
+ " \n",
+ " # Adding to the json\n",
+ " label_data[i] = {\n",
+ " \"name\":class_name,\n",
+ " \"coord\": (xmin, ymin, xmax, ymax),\n",
+ " \"img\":cropped_image\n",
+ " }\n",
+ "\n",
+ " datas[name] = {\n",
+ " \"img\" : image,\n",
+ " \"labels\" : label_data,\n",
+ " }\n",
+ " \n",
+ " return datas\n",
+ "\n",
+ "# Dict to convert str class name to int\n",
+ "name_to_int = {\n",
+ " \"danger\": 0,\n",
+ " \"interdiction\": 1,\n",
+ " \"obligation\": 2,\n",
+ " \"stop\": 3,\n",
+ " \"ceder\": 4,\n",
+ " \"frouge\": 5,\n",
+ " \"forange\": 6,\n",
+ " \"fvert\": 7,\n",
+ " \"ff\": 8,\n",
+ " \"empty\": 9\n",
+ "}"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "### 2) Fonction de création des datasets"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 3,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "def create_xy(datas):\n",
+ " # Creating arrays with all labels datas & classes\n",
+ " X = []\n",
+ " Y = []\n",
+ "\n",
+ " for name, data in datas.items():\n",
+ " for row in data[\"labels\"].values():\n",
+ " image_as_array = np.array(row[\"img\"]).flatten()\n",
+ " X.append(image_as_array)\n",
+ " Y.append(name_to_int[row[\"name\"]])\n",
+ "\n",
+ " X = np.array(X)\n",
+ " Y = np.array(Y)\n",
+ "\n",
+ " return X, Y"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "### 3) Création des datasets"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 4,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "# Training dataset\n",
+ "datas_train = load_data(\"../../data/train/images\", \"../../data/train/labels\")\n",
+ "X_train, Y_train = create_xy(datas=datas_train)\n",
+ "\n",
+ "# Validation dataset\n",
+ "datas_val = load_data(\"../../data/val/images\", \"../../data/val/labels\")\n",
+ "X_val, Y_val = create_xy(datas=datas_val)"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "### 4) Application de la méthode RandomTree"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 32,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "Taux d'erreur : 0.20168067226890757\n"
+ ]
+ }
+ ],
+ "source": [
+ "from sklearn.ensemble import RandomForestClassifier\n",
+ "\n",
+ "adaboost_clf = RandomForestClassifier(n_estimators=50) # To change\n",
+ "adaboost_clf.fit(X_train, Y_train)\n",
+ "y = adaboost_clf.predict(X_val)\n",
+ "\n",
+ "print(f\"Taux d'erreur : {np.mean(y != Y_val)}\")"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "### 5) Test de la méthode RandomTree avec application des caractéristiques HOG"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 6,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "from skimage.feature import hog\n",
+ "from skimage.color import rgb2gray\n",
+ "\n",
+ "def extract_hog(datas):\n",
+ " # Creating X array with all HOG information of images\n",
+ " X = []\n",
+ "\n",
+ " for name, data in datas.items():\n",
+ " for row in data[\"labels\"].values():\n",
+ " image_as_array = np.array(hog(rgb2gray(row[\"img\"]))).flatten()\n",
+ " X.append(image_as_array)\n",
+ "\n",
+ " return np.array(X)\n",
+ "\n",
+ "\n",
+ "# Update training dataset\n",
+ "X_train_HOG = extract_hog(datas=datas_train)\n",
+ "\n",
+ "# Update validation dataset\n",
+ "X_val_HOG = extract_hog(datas=datas_val)"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 7,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "Taux d'erreur : 0.2689075630252101\n"
+ ]
+ }
+ ],
+ "source": [
+ "adaboost_clf = RandomForestClassifier(n_estimators=10)\n",
+ "adaboost_clf.fit(X_train_HOG, Y_train)\n",
+ "y_HOG = adaboost_clf.predict(X_val_HOG)\n",
+ "\n",
+ "print(f\"Taux d'erreur : {np.mean(y_HOG != Y_val)}\")"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "### 6) Détermination du meilleur paramètre"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 33,
+ "metadata": {},
+ "outputs": [
+ {
+ "ename": "KeyboardInterrupt",
+ "evalue": "",
+ "output_type": "error",
+ "traceback": [
+ "\u001b[1;31m---------------------------------------------------------------------------\u001b[0m",
+ "\u001b[1;31mKeyboardInterrupt\u001b[0m Traceback (most recent call last)",
+ "Cell \u001b[1;32mIn[33], line 7\u001b[0m\n\u001b[0;32m 5\u001b[0m \u001b[38;5;28;01mfor\u001b[39;00m i \u001b[38;5;129;01min\u001b[39;00m \u001b[38;5;28mrange\u001b[39m(\u001b[38;5;241m1\u001b[39m, \u001b[38;5;241m100\u001b[39m, \u001b[38;5;241m10\u001b[39m):\n\u001b[0;32m 6\u001b[0m adaboost_clf \u001b[38;5;241m=\u001b[39m RandomForestClassifier(n_estimators\u001b[38;5;241m=\u001b[39mi) \u001b[38;5;66;03m# To change\u001b[39;00m\n\u001b[1;32m----> 7\u001b[0m \u001b[43madaboost_clf\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mfit\u001b[49m\u001b[43m(\u001b[49m\u001b[43mX_train\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mY_train\u001b[49m\u001b[43m)\u001b[49m\n\u001b[0;32m 8\u001b[0m y \u001b[38;5;241m=\u001b[39m adaboost_clf\u001b[38;5;241m.\u001b[39mpredict(X_val)\n\u001b[0;32m 9\u001b[0m tab\u001b[38;5;241m.\u001b[39mappend(np\u001b[38;5;241m.\u001b[39mmean(y \u001b[38;5;241m!=\u001b[39m Y_val))\n",
+ "File \u001b[1;32mc:\\Users\\victo\\AppData\\Local\\Programs\\Python\\Python312\\Lib\\site-packages\\sklearn\\base.py:1474\u001b[0m, in \u001b[0;36m_fit_context.<locals>.decorator.<locals>.wrapper\u001b[1;34m(estimator, *args, **kwargs)\u001b[0m\n\u001b[0;32m 1467\u001b[0m estimator\u001b[38;5;241m.\u001b[39m_validate_params()\n\u001b[0;32m 1469\u001b[0m \u001b[38;5;28;01mwith\u001b[39;00m config_context(\n\u001b[0;32m 1470\u001b[0m skip_parameter_validation\u001b[38;5;241m=\u001b[39m(\n\u001b[0;32m 1471\u001b[0m prefer_skip_nested_validation \u001b[38;5;129;01mor\u001b[39;00m global_skip_validation\n\u001b[0;32m 1472\u001b[0m )\n\u001b[0;32m 1473\u001b[0m ):\n\u001b[1;32m-> 1474\u001b[0m \u001b[38;5;28;01mreturn\u001b[39;00m \u001b[43mfit_method\u001b[49m\u001b[43m(\u001b[49m\u001b[43mestimator\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[38;5;241;43m*\u001b[39;49m\u001b[43margs\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[38;5;241;43m*\u001b[39;49m\u001b[38;5;241;43m*\u001b[39;49m\u001b[43mkwargs\u001b[49m\u001b[43m)\u001b[49m\n",
+ "File \u001b[1;32mc:\\Users\\victo\\AppData\\Local\\Programs\\Python\\Python312\\Lib\\site-packages\\sklearn\\ensemble\\_forest.py:489\u001b[0m, in \u001b[0;36mBaseForest.fit\u001b[1;34m(self, X, y, sample_weight)\u001b[0m\n\u001b[0;32m 478\u001b[0m trees \u001b[38;5;241m=\u001b[39m [\n\u001b[0;32m 479\u001b[0m \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39m_make_estimator(append\u001b[38;5;241m=\u001b[39m\u001b[38;5;28;01mFalse\u001b[39;00m, random_state\u001b[38;5;241m=\u001b[39mrandom_state)\n\u001b[0;32m 480\u001b[0m \u001b[38;5;28;01mfor\u001b[39;00m i \u001b[38;5;129;01min\u001b[39;00m \u001b[38;5;28mrange\u001b[39m(n_more_estimators)\n\u001b[0;32m 481\u001b[0m ]\n\u001b[0;32m 483\u001b[0m \u001b[38;5;66;03m# Parallel loop: we prefer the threading backend as the Cython code\u001b[39;00m\n\u001b[0;32m 484\u001b[0m \u001b[38;5;66;03m# for fitting the trees is internally releasing the Python GIL\u001b[39;00m\n\u001b[0;32m 485\u001b[0m \u001b[38;5;66;03m# making threading more efficient than multiprocessing in\u001b[39;00m\n\u001b[0;32m 486\u001b[0m \u001b[38;5;66;03m# that case. However, for joblib 0.12+ we respect any\u001b[39;00m\n\u001b[0;32m 487\u001b[0m \u001b[38;5;66;03m# parallel_backend contexts set at a higher level,\u001b[39;00m\n\u001b[0;32m 488\u001b[0m \u001b[38;5;66;03m# since correctness does not rely on using threads.\u001b[39;00m\n\u001b[1;32m--> 489\u001b[0m trees \u001b[38;5;241m=\u001b[39m \u001b[43mParallel\u001b[49m\u001b[43m(\u001b[49m\n\u001b[0;32m 490\u001b[0m \u001b[43m \u001b[49m\u001b[43mn_jobs\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[38;5;28;43mself\u001b[39;49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mn_jobs\u001b[49m\u001b[43m,\u001b[49m\n\u001b[0;32m 491\u001b[0m \u001b[43m \u001b[49m\u001b[43mverbose\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[38;5;28;43mself\u001b[39;49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mverbose\u001b[49m\u001b[43m,\u001b[49m\n\u001b[0;32m 492\u001b[0m \u001b[43m \u001b[49m\u001b[43mprefer\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[38;5;124;43m\"\u001b[39;49m\u001b[38;5;124;43mthreads\u001b[39;49m\u001b[38;5;124;43m\"\u001b[39;49m\u001b[43m,\u001b[49m\n\u001b[0;32m 493\u001b[0m \u001b[43m\u001b[49m\u001b[43m)\u001b[49m\u001b[43m(\u001b[49m\n\u001b[0;32m 494\u001b[0m \u001b[43m \u001b[49m\u001b[43mdelayed\u001b[49m\u001b[43m(\u001b[49m\u001b[43m_parallel_build_trees\u001b[49m\u001b[43m)\u001b[49m\u001b[43m(\u001b[49m\n\u001b[0;32m 495\u001b[0m \u001b[43m \u001b[49m\u001b[43mt\u001b[49m\u001b[43m,\u001b[49m\n\u001b[0;32m 496\u001b[0m \u001b[43m \u001b[49m\u001b[38;5;28;43mself\u001b[39;49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mbootstrap\u001b[49m\u001b[43m,\u001b[49m\n\u001b[0;32m 497\u001b[0m \u001b[43m \u001b[49m\u001b[43mX\u001b[49m\u001b[43m,\u001b[49m\n\u001b[0;32m 498\u001b[0m \u001b[43m \u001b[49m\u001b[43my\u001b[49m\u001b[43m,\u001b[49m\n\u001b[0;32m 499\u001b[0m \u001b[43m \u001b[49m\u001b[43msample_weight\u001b[49m\u001b[43m,\u001b[49m\n\u001b[0;32m 500\u001b[0m \u001b[43m \u001b[49m\u001b[43mi\u001b[49m\u001b[43m,\u001b[49m\n\u001b[0;32m 501\u001b[0m \u001b[43m \u001b[49m\u001b[38;5;28;43mlen\u001b[39;49m\u001b[43m(\u001b[49m\u001b[43mtrees\u001b[49m\u001b[43m)\u001b[49m\u001b[43m,\u001b[49m\n\u001b[0;32m 502\u001b[0m \u001b[43m \u001b[49m\u001b[43mverbose\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[38;5;28;43mself\u001b[39;49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mverbose\u001b[49m\u001b[43m,\u001b[49m\n\u001b[0;32m 503\u001b[0m \u001b[43m \u001b[49m\u001b[43mclass_weight\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[38;5;28;43mself\u001b[39;49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mclass_weight\u001b[49m\u001b[43m,\u001b[49m\n\u001b[0;32m 504\u001b[0m \u001b[43m \u001b[49m\u001b[43mn_samples_bootstrap\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[43mn_samples_bootstrap\u001b[49m\u001b[43m,\u001b[49m\n\u001b[0;32m 505\u001b[0m \u001b[43m \u001b[49m\u001b[43mmissing_values_in_feature_mask\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[43mmissing_values_in_feature_mask\u001b[49m\u001b[43m,\u001b[49m\n\u001b[0;32m 506\u001b[0m \u001b[43m \u001b[49m\u001b[43m)\u001b[49m\n\u001b[0;32m 507\u001b[0m \u001b[43m \u001b[49m\u001b[38;5;28;43;01mfor\u001b[39;49;00m\u001b[43m \u001b[49m\u001b[43mi\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mt\u001b[49m\u001b[43m \u001b[49m\u001b[38;5;129;43;01min\u001b[39;49;00m\u001b[43m \u001b[49m\u001b[38;5;28;43menumerate\u001b[39;49m\u001b[43m(\u001b[49m\u001b[43mtrees\u001b[49m\u001b[43m)\u001b[49m\n\u001b[0;32m 508\u001b[0m \u001b[43m\u001b[49m\u001b[43m)\u001b[49m\n\u001b[0;32m 510\u001b[0m \u001b[38;5;66;03m# Collect newly grown trees\u001b[39;00m\n\u001b[0;32m 511\u001b[0m \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39mestimators_\u001b[38;5;241m.\u001b[39mextend(trees)\n",
+ "File \u001b[1;32mc:\\Users\\victo\\AppData\\Local\\Programs\\Python\\Python312\\Lib\\site-packages\\sklearn\\utils\\parallel.py:67\u001b[0m, in \u001b[0;36mParallel.__call__\u001b[1;34m(self, iterable)\u001b[0m\n\u001b[0;32m 62\u001b[0m config \u001b[38;5;241m=\u001b[39m get_config()\n\u001b[0;32m 63\u001b[0m iterable_with_config \u001b[38;5;241m=\u001b[39m (\n\u001b[0;32m 64\u001b[0m (_with_config(delayed_func, config), args, kwargs)\n\u001b[0;32m 65\u001b[0m \u001b[38;5;28;01mfor\u001b[39;00m delayed_func, args, kwargs \u001b[38;5;129;01min\u001b[39;00m iterable\n\u001b[0;32m 66\u001b[0m )\n\u001b[1;32m---> 67\u001b[0m \u001b[38;5;28;01mreturn\u001b[39;00m \u001b[38;5;28;43msuper\u001b[39;49m\u001b[43m(\u001b[49m\u001b[43m)\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[38;5;21;43m__call__\u001b[39;49m\u001b[43m(\u001b[49m\u001b[43miterable_with_config\u001b[49m\u001b[43m)\u001b[49m\n",
+ "File \u001b[1;32mc:\\Users\\victo\\AppData\\Local\\Programs\\Python\\Python312\\Lib\\site-packages\\joblib\\parallel.py:1863\u001b[0m, in \u001b[0;36mParallel.__call__\u001b[1;34m(self, iterable)\u001b[0m\n\u001b[0;32m 1861\u001b[0m output \u001b[38;5;241m=\u001b[39m \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39m_get_sequential_output(iterable)\n\u001b[0;32m 1862\u001b[0m \u001b[38;5;28mnext\u001b[39m(output)\n\u001b[1;32m-> 1863\u001b[0m \u001b[38;5;28;01mreturn\u001b[39;00m output \u001b[38;5;28;01mif\u001b[39;00m \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39mreturn_generator \u001b[38;5;28;01melse\u001b[39;00m \u001b[38;5;28;43mlist\u001b[39;49m\u001b[43m(\u001b[49m\u001b[43moutput\u001b[49m\u001b[43m)\u001b[49m\n\u001b[0;32m 1865\u001b[0m \u001b[38;5;66;03m# Let's create an ID that uniquely identifies the current call. If the\u001b[39;00m\n\u001b[0;32m 1866\u001b[0m \u001b[38;5;66;03m# call is interrupted early and that the same instance is immediately\u001b[39;00m\n\u001b[0;32m 1867\u001b[0m \u001b[38;5;66;03m# re-used, this id will be used to prevent workers that were\u001b[39;00m\n\u001b[0;32m 1868\u001b[0m \u001b[38;5;66;03m# concurrently finalizing a task from the previous call to run the\u001b[39;00m\n\u001b[0;32m 1869\u001b[0m \u001b[38;5;66;03m# callback.\u001b[39;00m\n\u001b[0;32m 1870\u001b[0m \u001b[38;5;28;01mwith\u001b[39;00m \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39m_lock:\n",
+ "File \u001b[1;32mc:\\Users\\victo\\AppData\\Local\\Programs\\Python\\Python312\\Lib\\site-packages\\joblib\\parallel.py:1792\u001b[0m, in \u001b[0;36mParallel._get_sequential_output\u001b[1;34m(self, iterable)\u001b[0m\n\u001b[0;32m 1790\u001b[0m \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39mn_dispatched_batches \u001b[38;5;241m+\u001b[39m\u001b[38;5;241m=\u001b[39m \u001b[38;5;241m1\u001b[39m\n\u001b[0;32m 1791\u001b[0m \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39mn_dispatched_tasks \u001b[38;5;241m+\u001b[39m\u001b[38;5;241m=\u001b[39m \u001b[38;5;241m1\u001b[39m\n\u001b[1;32m-> 1792\u001b[0m res \u001b[38;5;241m=\u001b[39m \u001b[43mfunc\u001b[49m\u001b[43m(\u001b[49m\u001b[38;5;241;43m*\u001b[39;49m\u001b[43margs\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[38;5;241;43m*\u001b[39;49m\u001b[38;5;241;43m*\u001b[39;49m\u001b[43mkwargs\u001b[49m\u001b[43m)\u001b[49m\n\u001b[0;32m 1793\u001b[0m \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39mn_completed_tasks \u001b[38;5;241m+\u001b[39m\u001b[38;5;241m=\u001b[39m \u001b[38;5;241m1\u001b[39m\n\u001b[0;32m 1794\u001b[0m \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39mprint_progress()\n",
+ "File \u001b[1;32mc:\\Users\\victo\\AppData\\Local\\Programs\\Python\\Python312\\Lib\\site-packages\\sklearn\\utils\\parallel.py:129\u001b[0m, in \u001b[0;36m_FuncWrapper.__call__\u001b[1;34m(self, *args, **kwargs)\u001b[0m\n\u001b[0;32m 127\u001b[0m config \u001b[38;5;241m=\u001b[39m {}\n\u001b[0;32m 128\u001b[0m \u001b[38;5;28;01mwith\u001b[39;00m config_context(\u001b[38;5;241m*\u001b[39m\u001b[38;5;241m*\u001b[39mconfig):\n\u001b[1;32m--> 129\u001b[0m \u001b[38;5;28;01mreturn\u001b[39;00m \u001b[38;5;28;43mself\u001b[39;49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mfunction\u001b[49m\u001b[43m(\u001b[49m\u001b[38;5;241;43m*\u001b[39;49m\u001b[43margs\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[38;5;241;43m*\u001b[39;49m\u001b[38;5;241;43m*\u001b[39;49m\u001b[43mkwargs\u001b[49m\u001b[43m)\u001b[49m\n",
+ "File \u001b[1;32mc:\\Users\\victo\\AppData\\Local\\Programs\\Python\\Python312\\Lib\\site-packages\\sklearn\\ensemble\\_forest.py:192\u001b[0m, in \u001b[0;36m_parallel_build_trees\u001b[1;34m(tree, bootstrap, X, y, sample_weight, tree_idx, n_trees, verbose, class_weight, n_samples_bootstrap, missing_values_in_feature_mask)\u001b[0m\n\u001b[0;32m 189\u001b[0m \u001b[38;5;28;01melif\u001b[39;00m class_weight \u001b[38;5;241m==\u001b[39m \u001b[38;5;124m\"\u001b[39m\u001b[38;5;124mbalanced_subsample\u001b[39m\u001b[38;5;124m\"\u001b[39m:\n\u001b[0;32m 190\u001b[0m curr_sample_weight \u001b[38;5;241m*\u001b[39m\u001b[38;5;241m=\u001b[39m compute_sample_weight(\u001b[38;5;124m\"\u001b[39m\u001b[38;5;124mbalanced\u001b[39m\u001b[38;5;124m\"\u001b[39m, y, indices\u001b[38;5;241m=\u001b[39mindices)\n\u001b[1;32m--> 192\u001b[0m \u001b[43mtree\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43m_fit\u001b[49m\u001b[43m(\u001b[49m\n\u001b[0;32m 193\u001b[0m \u001b[43m \u001b[49m\u001b[43mX\u001b[49m\u001b[43m,\u001b[49m\n\u001b[0;32m 194\u001b[0m \u001b[43m \u001b[49m\u001b[43my\u001b[49m\u001b[43m,\u001b[49m\n\u001b[0;32m 195\u001b[0m \u001b[43m \u001b[49m\u001b[43msample_weight\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[43mcurr_sample_weight\u001b[49m\u001b[43m,\u001b[49m\n\u001b[0;32m 196\u001b[0m \u001b[43m \u001b[49m\u001b[43mcheck_input\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[38;5;28;43;01mFalse\u001b[39;49;00m\u001b[43m,\u001b[49m\n\u001b[0;32m 197\u001b[0m \u001b[43m \u001b[49m\u001b[43mmissing_values_in_feature_mask\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[43mmissing_values_in_feature_mask\u001b[49m\u001b[43m,\u001b[49m\n\u001b[0;32m 198\u001b[0m \u001b[43m \u001b[49m\u001b[43m)\u001b[49m\n\u001b[0;32m 199\u001b[0m \u001b[38;5;28;01melse\u001b[39;00m:\n\u001b[0;32m 200\u001b[0m tree\u001b[38;5;241m.\u001b[39m_fit(\n\u001b[0;32m 201\u001b[0m X,\n\u001b[0;32m 202\u001b[0m y,\n\u001b[1;32m (...)\u001b[0m\n\u001b[0;32m 205\u001b[0m missing_values_in_feature_mask\u001b[38;5;241m=\u001b[39mmissing_values_in_feature_mask,\n\u001b[0;32m 206\u001b[0m )\n",
+ "File \u001b[1;32mc:\\Users\\victo\\AppData\\Local\\Programs\\Python\\Python312\\Lib\\site-packages\\sklearn\\tree\\_classes.py:472\u001b[0m, in \u001b[0;36mBaseDecisionTree._fit\u001b[1;34m(self, X, y, sample_weight, check_input, missing_values_in_feature_mask)\u001b[0m\n\u001b[0;32m 461\u001b[0m \u001b[38;5;28;01melse\u001b[39;00m:\n\u001b[0;32m 462\u001b[0m builder \u001b[38;5;241m=\u001b[39m BestFirstTreeBuilder(\n\u001b[0;32m 463\u001b[0m splitter,\n\u001b[0;32m 464\u001b[0m min_samples_split,\n\u001b[1;32m (...)\u001b[0m\n\u001b[0;32m 469\u001b[0m \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39mmin_impurity_decrease,\n\u001b[0;32m 470\u001b[0m )\n\u001b[1;32m--> 472\u001b[0m \u001b[43mbuilder\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mbuild\u001b[49m\u001b[43m(\u001b[49m\u001b[38;5;28;43mself\u001b[39;49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mtree_\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mX\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43my\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43msample_weight\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mmissing_values_in_feature_mask\u001b[49m\u001b[43m)\u001b[49m\n\u001b[0;32m 474\u001b[0m \u001b[38;5;28;01mif\u001b[39;00m \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39mn_outputs_ \u001b[38;5;241m==\u001b[39m \u001b[38;5;241m1\u001b[39m \u001b[38;5;129;01mand\u001b[39;00m is_classifier(\u001b[38;5;28mself\u001b[39m):\n\u001b[0;32m 475\u001b[0m \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39mn_classes_ \u001b[38;5;241m=\u001b[39m \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39mn_classes_[\u001b[38;5;241m0\u001b[39m]\n",
+ "\u001b[1;31mKeyboardInterrupt\u001b[0m: "
+ ]
+ }
+ ],
+ "source": [
+ "import matplotlib.pyplot as plt\n",
+ "\n",
+ "tab = []\n",
+ "\n",
+ "for i in range(1, 100, 10):\n",
+ " adaboost_clf = RandomForestClassifier(n_estimators=i) # To change\n",
+ " adaboost_clf.fit(X_train, Y_train)\n",
+ " y = adaboost_clf.predict(X_val)\n",
+ " tab.append(np.mean(y != Y_val))\n",
+ "\n",
+ "plt.bar(range(len(tab)), tab, color='skyblue')\n",
+ "plt.show()\n",
+ "\n",
+ "print(tab)\n",
+ "print(f\"Taux d'erreur : {np.mean(y != Y_val)}\")"
+ ]
+ }
+ ],
+ "metadata": {
+ "kernelspec": {
+ "display_name": "venv",
+ "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": "3.12.2"
+ }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 2
+}
diff --git a/machine_learning/classification/SVM.ipynb b/machine_learning/classification/SVM.ipynb
new file mode 100644
index 0000000000000000000000000000000000000000..26ee1cc09bf7e8d99ea6665d97e6e3796f3a97b3
--- /dev/null
+++ b/machine_learning/classification/SVM.ipynb
@@ -0,0 +1,369 @@
+{
+ "cells": [
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "## Entrainement d'un modèle avec la méthode des SVM"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 20,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "import os\n",
+ "import numpy as np\n",
+ "import random\n",
+ "from PIL import Image"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "### 1) Fonctions de Preprocessing des datasets"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 21,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "AVERAGE_SIZE_IMAGE = (127, 145) # Thanks to the stats, we know that size of bbox will be (127, 145) -> Average size of labels \n",
+ "\n",
+ "def generate_empty_bbox(image_width, image_height):\n",
+ " \"\"\" \n",
+ " Generate an empty box for images without label\n",
+ " \"\"\"\n",
+ " # Generating random coords for the bbox\n",
+ " x_min = random.randint(0, image_width - AVERAGE_SIZE_IMAGE[0])\n",
+ " y_min = random.randint(0, image_height - AVERAGE_SIZE_IMAGE[1])\n",
+ " \n",
+ " # Compute complete coords of the bbox\n",
+ " x_max = x_min + AVERAGE_SIZE_IMAGE[0]\n",
+ " y_max = y_min + AVERAGE_SIZE_IMAGE[1]\n",
+ " \n",
+ " return (x_min, y_min, x_max, y_max)\n",
+ "\n",
+ "def load_data(image_dir, label_dir):\n",
+ " \"\"\" \n",
+ " Create a dict with all the usefull datas of the dataset\n",
+ " datas = {\n",
+ " \"XXXX\" (name of the file) : {\n",
+ " \"img\" : image as an array,\n",
+ " \"labels\" (data of the labels): {\n",
+ " \"X\" index of the label (0,1,...,n) : {\n",
+ " \"name\" : name of the label,\n",
+ " \"coord\" : coord of the label like xmin, ymin, xmax, ymax,\n",
+ " \"img\" : crooped img of the label,\n",
+ " }\n",
+ " }\n",
+ " }\n",
+ " }\n",
+ " \"\"\"\n",
+ " \n",
+ " datas = {}\n",
+ "\n",
+ " for image_file in os.listdir(image_dir):\n",
+ " # Computing name and files paths\n",
+ " image_path = image_dir + '/' + image_file\n",
+ " name = image_file.split('.')[0]\n",
+ " label_path = label_dir + '/' + name + '.csv'\n",
+ " \n",
+ " # Import image as array\n",
+ " image = np.array(Image.open(image_path))\n",
+ "\n",
+ " # Import labels as array \n",
+ " with open(label_path, 'r') as file:\n",
+ " rows = file.readlines()\n",
+ "\n",
+ " label_data = {}\n",
+ " if rows == ['\\n']: # Create a random empty label to balance model\n",
+ " # Create random coords for empty label\n",
+ " xmin, ymin, xmax, ymax = generate_empty_bbox(image.shape[1], image.shape[0])\n",
+ " \n",
+ " # Get the cropped image (as array) of the label\n",
+ " cropped_image = np.array(Image.fromarray(image[ymin:ymax, xmin:xmax]).resize(AVERAGE_SIZE_IMAGE))\n",
+ " \n",
+ " label_data[0] = {\n",
+ " \"name\":\"empty\",\n",
+ " \"coord\": (xmin, ymin, xmax, ymax),\n",
+ " \"img\":cropped_image\n",
+ " }\n",
+ " else:\n",
+ " for i, row in enumerate(rows): # One image can contain several labels\n",
+ " row = row.strip().split(\",\")\n",
+ "\n",
+ " # Compute coords of the label\n",
+ " xmin, ymin, xmax, ymax = map(int, row[0:4])\n",
+ "\n",
+ " # Get the label name\n",
+ " class_name = row[4]\n",
+ "\n",
+ " # Get the cropped image (as array) of the label\n",
+ " cropped_image = np.array(Image.fromarray(image[ymin:ymax, xmin:xmax]).resize(AVERAGE_SIZE_IMAGE))\n",
+ " \n",
+ " # Adding to the json\n",
+ " label_data[i] = {\n",
+ " \"name\":class_name,\n",
+ " \"coord\": (xmin, ymin, xmax, ymax),\n",
+ " \"img\":cropped_image\n",
+ " }\n",
+ "\n",
+ " datas[name] = {\n",
+ " \"img\" : image,\n",
+ " \"labels\" : label_data,\n",
+ " }\n",
+ " \n",
+ " return datas\n",
+ "\n",
+ "# Dict to convert str class name to int\n",
+ "name_to_int = {\n",
+ " \"danger\": 0,\n",
+ " \"interdiction\": 1,\n",
+ " \"obligation\": 2,\n",
+ " \"stop\": 3,\n",
+ " \"ceder\": 4,\n",
+ " \"frouge\": 5,\n",
+ " \"forange\": 6,\n",
+ " \"fvert\": 7,\n",
+ " \"ff\": 8,\n",
+ " \"empty\": 9\n",
+ "}"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "### 2) Fonction de création des datasets"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 22,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "def create_xy(datas):\n",
+ " # Creating arrays with all labels datas & classes\n",
+ " X = []\n",
+ " Y = []\n",
+ "\n",
+ " for name, data in datas.items():\n",
+ " for row in data[\"labels\"].values():\n",
+ " image_as_array = np.array(row[\"img\"]).flatten()\n",
+ " X.append(image_as_array)\n",
+ " Y.append(name_to_int[row[\"name\"]])\n",
+ "\n",
+ " X = np.array(X)\n",
+ " Y = np.array(Y)\n",
+ "\n",
+ " return X, Y"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "### 3) Création des datasets"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 23,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "# Training dataset\n",
+ "datas_train = load_data(\"../../data/train/images\", \"../../data/train/labels\")\n",
+ "X_train, Y_train = create_xy(datas=datas_train)\n",
+ "\n",
+ "# Validation dataset\n",
+ "datas_val = load_data(\"../../data/val/images\", \"../../data/val/labels\")\n",
+ "X_val, Y_val = create_xy(datas=datas_val)"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "### 4) Application de la méthode des SVM"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 24,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "Taux d'erreur : 0.226890756302521\n"
+ ]
+ }
+ ],
+ "source": [
+ "from sklearn import svm\n",
+ "\n",
+ "svm_model = svm.SVC(kernel='linear') \n",
+ "svm_model.fit(X_train, Y_train)\n",
+ "y = svm_model.predict(X_val)\n",
+ "\n",
+ "print(f\"Taux d'erreur : {np.mean(y != Y_val)}\")"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "### 5) Test de la méthode SVM avec application des caractéristiques HOG"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 27,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "from skimage.feature import hog\n",
+ "from skimage.color import rgb2gray\n",
+ "\n",
+ "def extract_hog(datas):\n",
+ " # Creating X array with all HOG information of images\n",
+ " X = []\n",
+ "\n",
+ " for name, data in datas.items():\n",
+ " for row in data[\"labels\"].values():\n",
+ " image_as_array = np.array(hog(rgb2gray(row[\"img\"]))).flatten()\n",
+ " X.append(image_as_array)\n",
+ "\n",
+ " return np.array(X)\n",
+ "\n",
+ "\n",
+ "# Update training dataset\n",
+ "X_train_HOG = extract_hog(datas=datas_train)\n",
+ "\n",
+ "# Update validation dataset\n",
+ "X_val_HOG = extract_hog(datas=datas_val)"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 28,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "Taux d'erreur : 0.15966386554621848\n"
+ ]
+ }
+ ],
+ "source": [
+ "svm_model = svm.SVC(kernel='linear') \n",
+ "svm_model.fit(X_train_HOG, Y_train)\n",
+ "y_HOG = svm_model.predict(X_val_HOG)\n",
+ "\n",
+ "print(f\"Taux d'erreur : {np.mean(y_HOG != Y_val)}\")"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "### 6) Test de la méthode SVM avec application des LPB"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 34,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stderr",
+ "output_type": "stream",
+ "text": [
+ "c:\\Users\\victo\\AppData\\Local\\Programs\\Python\\Python312\\Lib\\site-packages\\skimage\\feature\\texture.py:360: UserWarning: Applying `local_binary_pattern` to floating-point images may give unexpected results when small numerical differences between adjacent pixels are present. It is recommended to use this function with images of integer dtype.\n",
+ " warnings.warn(\n"
+ ]
+ },
+ {
+ "ename": "ValueError",
+ "evalue": "setting an array element with a sequence. The requested array has an inhomogeneous shape after 1 dimensions. The detected shape was (1071,) + inhomogeneous part.",
+ "output_type": "error",
+ "traceback": [
+ "\u001b[1;31m---------------------------------------------------------------------------\u001b[0m",
+ "\u001b[1;31mValueError\u001b[0m Traceback (most recent call last)",
+ "Cell \u001b[1;32mIn[34], line 17\u001b[0m\n\u001b[0;32m 13\u001b[0m \u001b[38;5;28;01mreturn\u001b[39;00m np\u001b[38;5;241m.\u001b[39marray(X)\n\u001b[0;32m 16\u001b[0m \u001b[38;5;66;03m# Update training dataset\u001b[39;00m\n\u001b[1;32m---> 17\u001b[0m X_train_LBP \u001b[38;5;241m=\u001b[39m \u001b[43mextract_LBP\u001b[49m\u001b[43m(\u001b[49m\u001b[43mdatas\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[43mdatas_train\u001b[49m\u001b[43m)\u001b[49m\n\u001b[0;32m 19\u001b[0m \u001b[38;5;66;03m# Update validation dataset\u001b[39;00m\n\u001b[0;32m 20\u001b[0m X_val_LBP \u001b[38;5;241m=\u001b[39m extract_LBP(datas\u001b[38;5;241m=\u001b[39mdatas_val)\n",
+ "Cell \u001b[1;32mIn[34], line 13\u001b[0m, in \u001b[0;36mextract_LBP\u001b[1;34m(datas)\u001b[0m\n\u001b[0;32m 10\u001b[0m image_as_array \u001b[38;5;241m=\u001b[39m np\u001b[38;5;241m.\u001b[39marray(hog(local_binary_pattern(rgb2gray(data[\u001b[38;5;124m\"\u001b[39m\u001b[38;5;124mimg\u001b[39m\u001b[38;5;124m\"\u001b[39m]), P \u001b[38;5;241m=\u001b[39m \u001b[38;5;241m8\u001b[39m, R \u001b[38;5;241m=\u001b[39m \u001b[38;5;241m1\u001b[39m)))\u001b[38;5;241m.\u001b[39mflatten()\n\u001b[0;32m 11\u001b[0m X\u001b[38;5;241m.\u001b[39mappend(image_as_array)\n\u001b[1;32m---> 13\u001b[0m \u001b[38;5;28;01mreturn\u001b[39;00m \u001b[43mnp\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43marray\u001b[49m\u001b[43m(\u001b[49m\u001b[43mX\u001b[49m\u001b[43m)\u001b[49m\n",
+ "\u001b[1;31mValueError\u001b[0m: setting an array element with a sequence. The requested array has an inhomogeneous shape after 1 dimensions. The detected shape was (1071,) + inhomogeneous part."
+ ]
+ }
+ ],
+ "source": [
+ "import cv2\n",
+ "\n",
+ "def extract_SIFT(datas):\n",
+ " # Creating X array with all HOG information of images\n",
+ " X = []\n",
+ " sift = cv2.SIFT_create()\n",
+ "\n",
+ " for name, data in datas.items():\n",
+ " for row in data[\"labels\"].values():\n",
+ " gray_image = cv2.cvtColor(data[\"img\"], cv2.COLOR_RGB2GRAY)\n",
+ " keypoints, descriptors = sift.detectAndCompute(gray_image, None)\n",
+ " if descriptors is not None:\n",
+ " X.append(descriptors.flatten())\n",
+ "\n",
+ " return np.array(X)\n",
+ "\n",
+ "\n",
+ "# Update training dataset\n",
+ "X_train_LBP = extract_SIFT(datas=datas_train)\n",
+ "\n",
+ "# Update validation dataset\n",
+ "X_val_LBP = extract_SIFT(datas=datas_val)"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": null,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "svm_model = svm.SVC(kernel='linear') \n",
+ "svm_model.fit(X_train_LBP, Y_train)\n",
+ "y_LBP = svm_model.predict(X_val_LBP)\n",
+ "\n",
+ "print(f\"Taux d'erreur : {np.mean(y_LBP != Y_val)}\")"
+ ]
+ }
+ ],
+ "metadata": {
+ "kernelspec": {
+ "display_name": "venv",
+ "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": "3.12.2"
+ }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 2
+}
diff --git a/machine_learning/learn.ipynb b/machine_learning/learn.ipynb
deleted file mode 100644
index 18386e8cd90048a8b7bc3cdf9b5bc8d8a71aa9df..0000000000000000000000000000000000000000
--- a/machine_learning/learn.ipynb
+++ /dev/null
@@ -1,229 +0,0 @@
-{
- "cells": [
- {
- "cell_type": "markdown",
- "metadata": {},
- "source": [
- "## Entrainement d'un modèle avec la méthode des SVM"
- ]
- },
- {
- "cell_type": "markdown",
- "metadata": {},
- "source": [
- "#### Chargement des données d'entrainement"
- ]
- },
- {
- "cell_type": "code",
- "execution_count": 1,
- "metadata": {},
- "outputs": [],
- "source": [
- "import os\n",
- "import cv2\n",
- "import numpy as np\n",
- "import random"
- ]
- },
- {
- "cell_type": "code",
- "execution_count": 2,
- "metadata": {},
- "outputs": [],
- "source": [
- "\"\"\" \n",
- "We will create a dict with all the usefull datas of the training dataset\n",
- "datas = {\n",
- " \"XXXX\" (name of the file) : {\n",
- " \"img\" : ndarray of the image,\n",
- " \"labels\" (data of the labels): {\n",
- " \"X\" index of the label (0,1,...,n) : {\n",
- " \"name\" : name of the label,\n",
- " \"coord\" : coord of the label like xmin, ymin, xmax, ymax,\n",
- " \"img\" : crooped img of the label,\n",
- " }\n",
- " }\n",
- " }\n",
- "}\n",
- "\n",
- "\"\"\"\n",
- "\n",
- "def generate_empty_bbox(image_width, image_height):\n",
- " # Thanks to the stats, we know that size of bbox will be (127, 145) -> Average size of labels \n",
- " # Génération de coordonnées aléatoires pour le coin supérieur gauche de la boundebox\n",
- " x_min = random.randint(0, image_width - 127)\n",
- " y_min = random.randint(0, image_height - 145)\n",
- " \n",
- " # Calcul des coordonnées du coin inférieur droit de la boundebox\n",
- " x_max = x_min + 127\n",
- " y_max = y_min + 145\n",
- " \n",
- " return (x_min, y_min, x_max, y_max)\n",
- "\n",
- "def load_data(image_dir, label_dir):\n",
- " datas = {}\n",
- "\n",
- " for image_file in os.listdir(image_dir):\n",
- " # Computing name and files paths\n",
- " image_path = image_dir + '/' + image_file\n",
- " name = image_file.split('.')[0]\n",
- " label_path = label_dir + '/' + name + '.csv'\n",
- " \n",
- " # Import image as array\n",
- " image = cv2.imread(image_path)\n",
- "\n",
- " # Import labels as array \n",
- " with open(label_path, 'r') as file:\n",
- " rows = file.readlines()\n",
- "\n",
- " label_data = {}\n",
- " if rows == ['\\n']: # Create a random empty label to balance model\n",
- " # Create random coords for empty label\n",
- " xmin, ymin, xmax, ymax = generate_empty_bbox(image.shape[1], image.shape[0])\n",
- " \n",
- " # Get the cropped image (as array) of the label\n",
- " cropped_image = image[ymin:ymax, xmin:xmax]\n",
- " \n",
- " label_data[0] = {\n",
- " \"name\":\"empty\",\n",
- " \"coord\": (xmin, ymin, xmax, ymax),\n",
- " \"img\":cropped_image\n",
- " }\n",
- " else:\n",
- " for i, row in enumerate(rows): # One image can contain several labels\n",
- " row = row.strip().split(\",\")\n",
- "\n",
- " # Compute coords of the label\n",
- " xmin, ymin, xmax, ymax = map(int, row[0:4])\n",
- "\n",
- " # Get the label name\n",
- " class_name = row[4]\n",
- "\n",
- " # Get the cropped image (as array) of the label\n",
- " cropped_image = image[ymin:ymax, xmin:xmax]\n",
- " \n",
- " # Adding to the json\n",
- " label_data[i] = {\n",
- " \"name\":class_name,\n",
- " \"coord\": (xmin, ymin, xmax, ymax),\n",
- " \"img\":cropped_image\n",
- " }\n",
- "\n",
- " datas[name] = {\n",
- " \"img\" : image,\n",
- " \"labels\" : label_data,\n",
- " }\n",
- " \n",
- " return datas"
- ]
- },
- {
- "cell_type": "code",
- "execution_count": 3,
- "metadata": {},
- "outputs": [],
- "source": [
- "# Creating the dict of the datas \n",
- "\n",
- "datas = load_data(\"../data/train/images\", \"../data/train/labels\")"
- ]
- },
- {
- "cell_type": "code",
- "execution_count": 4,
- "metadata": {},
- "outputs": [],
- "source": [
- "def extract_features(img):\n",
- " # Convertion to gray level\n",
- " gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)\n",
- "\n",
- " # Color Hist\n",
- " hist_color = cv2.calcHist([img], [0, 1, 2], None, [8, 8, 8], [0, 256, 0, 256, 0, 256])\n",
- " hist_color = cv2.normalize(hist_color, hist_color).flatten()\n",
- " \n",
- " # Gradient Hist\n",
- " sobelx = cv2.Sobel(gray, cv2.CV_64F, 1, 0, ksize=5)\n",
- " sobely = cv2.Sobel(gray, cv2.CV_64F, 0, 1, ksize=5)\n",
- " grad_mag = np.sqrt(sobelx**2 + sobely**2)\n",
- " hist_gradient = cv2.calcHist([grad_mag.astype(np.uint8)], [0], None, [16], [0, 256])\n",
- " hist_gradient = cv2.normalize(hist_gradient, hist_gradient).flatten()\n",
- " \n",
- " return np.concatenate((hist_color, hist_gradient))\n",
- "\n",
- "\n",
- "# Dict to convert str class name to int\n",
- "name_to_int = {\n",
- " \"danger\": 0,\n",
- " \"interdiction\": 1,\n",
- " \"obligation\": 2,\n",
- " \"stop\": 3,\n",
- " \"ceder\": 4,\n",
- " \"frouge\": 5,\n",
- " \"forange\": 6,\n",
- " \"fvert\": 7,\n",
- " \"ff\": 8,\n",
- " \"empty\": 9\n",
- "}\n",
- "\n",
- "\n",
- "# Creating arrays with all labels datas & classes\n",
- "X_train = []\n",
- "Y_train = []\n",
- "\n",
- "for name, data in datas.items():\n",
- " for row in data[\"labels\"].values():\n",
- " X_train.append(extract_features(row[\"img\"]))\n",
- " Y_train.append(name_to_int[row[\"name\"]])\n",
- "\n",
- "X_train = np.array(X_train)\n",
- "Y_train = np.array(Y_train)"
- ]
- },
- {
- "cell_type": "code",
- "execution_count": 6,
- "metadata": {},
- "outputs": [
- {
- "name": "stdout",
- "output_type": "stream",
- "text": [
- "[1 1 0 ... 1 5 7]\n"
- ]
- }
- ],
- "source": [
- "from sklearn import svm\n",
- "from sklearn.metrics import accuracy_score, precision_score, recall_score, f1_score\n",
- "\n",
- "svm_model = svm.SVC(kernel='linear') # Choix du noyau linéaire\n",
- "svm_model.fit(X_train, Y_train)\n",
- "\n",
- "print(svm_model)"
- ]
- }
- ],
- "metadata": {
- "kernelspec": {
- "display_name": "venv",
- "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
-}