Updating pointnet_test, processing_data_2

RoboticsLabURJC · Dec 23, 2024 · aef744b · aef744b
1 parent 52b0cbb
commit aef744b
Show file tree

Hide file tree

Showing 2 changed files with 142 additions and 470 deletions.
diff --git a/segmentation/pointnet_test.ipynb b/segmentation/pointnet_test.ipynb
@@ -4,7 +4,14 @@
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "## Implementación PointNet en Keras, Tensorflow 2.0"
+    "# Implementación PointNet en Keras, Tensorflow 2.0"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "## Lectura y preparación del conjunto de datos"
    ]
   },
   {
@@ -27,16 +34,32 @@
     "from sklearn.model_selection import train_test_split\n"
    ]
   },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "### Cargar y Leer archivos _.csv_ de cada subconjunto"
+   ]
+  },
   {
    "cell_type": "code",
-   "execution_count": 15,
+   "execution_count": null,
    "metadata": {},
    "outputs": [],
    "source": [
-    "path_x_train = \"/home/felix/Escritorio/TFG/datasets_norm/goose_norm_test/x_train/\"\n",
-    "path_y_train = \"/home/felix/Escritorio/TFG/datasets_norm/goose_norm_test/y_train/\"\n",
-    "path_x_val = \"/home/felix/Escritorio/TFG/datasets_norm/goose_norm_test/x_val/\"\n",
-    "path_y_val = \"/home/felix/Escritorio/TFG/datasets_norm/goose_norm_test/y_val/\"\n",
+    "# Linux\n",
+    "\n",
+    "#path_x_train = \"/home/felix/Escritorio/TFG/datasets_norm/goose_norm_test/x_train/\"\n",
+    "#path_y_train = \"/home/felix/Escritorio/TFG/datasets_norm/goose_norm_test/y_train/\"\n",
+    "#path_x_val = \"/home/felix/Escritorio/TFG/datasets_norm/goose_norm_test/x_val/\"\n",
+    "#path_y_val = \"/home/felix/Escritorio/TFG/datasets_norm/goose_norm_test/y_val/\"\n",
+    "\n",
+    "# MacOS\n",
+    "\n",
+    "path_x_train = \"/Users/felixmaral/Desktop/TFG/dataset_norm/goose_teesting/x_train/\"\n",
+    "path_y_train = \"/Users/felixmaral/Desktop/TFG/dataset_norm/goose_teesting/y_train/\"\n",
+    "path_x_val = \"/Users/felixmaral/Desktop/TFG/dataset_norm/goose_teesting/x_val/\"\n",
+    "path_y_val = \"/Users/felixmaral/Desktop/TFG/dataset_norm/goose_teesting/y_val/\"\n",
     "\n",
     "file_list_x_train = sorted(os.listdir(path_x_train))\n",
     "file_list_y_train = sorted(os.listdir(path_y_train))\n",
@@ -49,47 +72,119 @@
     "Y_val = []\n",
     "\n",
     "for file in file_list_x_train:\n",
-    "    df = pd.read_csv(path_x_train+file)\n",
-    "    X_train.append(df)\n",
+    "    df_x_train = pd.read_csv(path_x_train+file)\n",
+    "    X_train.append(df_x_train)\n",
     "\n",
     "for file in file_list_y_train:\n",
-    "    df['sem_label'] = pd.read_csv(path_y_train+file)\n",
-    "    Y_train.append(df)\n",
+    "    df_y_train = pd.DataFrame()\n",
+    "    df_y_train['sem_label'] = pd.read_csv(path_y_train+file)\n",
+    "    Y_train.append(df_y_train)\n",
     "\n",
     "for file in file_list_x_val:\n",
-    "    df = pd.read_csv(path_x_val+file)\n",
-    "    X_val.append(df)\n",
+    "    df_x_val = pd.read_csv(path_x_val+file)\n",
+    "    X_val.append(df_x_val)\n",
     "\n",
     "for file in file_list_y_val:\n",
-    "    df['sem_label'] = pd.read_csv(path_y_val+file)\n",
-    "    Y_val.append(df)"
+    "    df_y_val = pd.DataFrame()\n",
+    "    df_y_val['sem_label'] = pd.read_csv(path_y_val+file)\n",
+    "    Y_val.append(df_y_val)"
    ]
   },
   {
    "cell_type": "code",
-   "execution_count": 19,
+   "execution_count": null,
    "metadata": {},
-   "outputs": [
-    {
-     "name": "stdout",
-     "output_type": "stream",
-     "text": [
-      "[50 41 51 59 17 38 23 31  4  8 47  0 55 45]\n",
-      "[50 59 51 27 17 41  8 28  4 31  0]\n"
-     ]
-    }
-   ],
+   "outputs": [],
    "source": [
     "# Estudio de clases\n",
     "\n",
     "Y_train_concat = pd.concat(Y_train)\n",
     "Y_val_concat = pd.concat(Y_val)\n",
+    "Y_concat = pd.concat(([pd.DataFrame(Y_train_concat), pd.DataFrame(Y_val_concat)]))\n",
     "\n",
     "clases_unicas_train = Y_train_concat['sem_label'].unique()\n",
     "clases_unicas_val = Y_val_concat['sem_label'].unique()\n",
+    "clases_unicas = set([int(x) for x in np.concatenate((clases_unicas_train, clases_unicas_val))])\n",
+    "\n",
+    "print(\"Clases definidas en el conjunto de entrenamiento: \", clases_unicas_train)\n",
+    "print(\"\\nClases definidas en el conjunto de validación: \",clases_unicas_val)\n",
+    "print(\"\\nClases definidas en el conjunto de datos: \",clases_unicas)\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# Calcula las clases faltantes en Xtrain\n",
+    "clases_faltantes_train = list(set(clases_unicas_train) - set(clases_unicas_val))\n",
+    "clases_unicas_df = np.concatenate((clases_unicas_train, clases_faltantes_train))\n",
+    "\n",
+    "# Calcula las clases existentes en el conjunto de datos\n",
+    "clases_unicas_df = sorted(set(clases_unicas_df))\n",
+    "\n",
+    "print(\"En el conjunto de datos hay: \", len(clases_unicas), \"clases únicas\")\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# Histograma de frecuencia de clases\n",
+    "\n",
+    "from matplotlib.ticker import FuncFormatter\n",
+    "\n",
+    "# Crear el histograma\n",
+    "plt.figure(figsize=(16, 6))\n",
+    "plt.style.use('ggplot')\n",
+    "plt.hist(Y_concat['sem_label'], \n",
+    "         bins=np.arange(Y_concat['sem_label'].min(), Y_concat['sem_label'].max() + 2), \n",
+    "         edgecolor='k', \n",
+    "         alpha=1, \n",
+    "         align='mid')\n",
+    "\n",
+    "# Chatgpt code\n",
+    "# Configurar el formateador para el eje Y\n",
+    "#ax = plt.gca()\n",
+    "#formatter = FuncFormatter(lambda x, _: f\"{int(x):,}\")  # Formato con separadores de miles\n",
+    "#ax.yaxis.set_major_formatter(formatter)  # Aplicar el formateador\n",
+    "\n",
+    "# Personalización del gráfico\n",
+    "plt.title(\"Histograma de Etiquetas Semánticas\", pad=20)\n",
+    "plt.xlabel(\"Etiqueta Semántica\", labelpad=10)\n",
+    "plt.ylabel(\"Frecuencia\", labelpad=10)\n",
+    "plt.grid(axis='y', linestyle='--', alpha=0.7)\n",
+    "\n",
+    "# Configurar las marcas del eje X en incrementos de 1\n",
+    "x_ticks = np.arange(Y_concat['sem_label'].min(), Y_concat['sem_label'].max() + 2, 1)\n",
+    "plt.xticks(x_ticks, fontsize=8)\n",
+    "\n",
+    "# Ajustar padding de los x_ticks\n",
+    "plt.gca().tick_params(axis='x', pad=10)  # Aumenta el espacio entre los ticks y el eje\n",
+    "\n",
+    "# Mostrar el gráfico\n",
+    "plt.show()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# Clases que faltan en el conjunto de datos\n",
+    "\n",
+    "GOOSE_CLASSES = [i for i in range(0, 64)]\n",
+    "\n",
+    "# Calcular clases faltantes\n",
+    "\n",
+    "clases_faltantes = [i for i in GOOSE_CLASSES if i not in clases_unicas]\n",
     "\n",
-    "print(clases_unicas_train)\n",
-    "print(clases_unicas_val)\n"
+    "print(f\"Hay {len(clases_faltantes)} clases que no están representadas en este conjunto de datos: \")\n",
+    "print(clases_faltantes)\n"
    ]
   },
   {
@@ -102,7 +197,7 @@
  ],
  "metadata": {
   "kernelspec": {
-   "display_name": "usr",
+   "display_name": "tfg",
    "language": "python",
    "name": "python3"
   },