185 lines (184 with data), 4.3 kB

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "from sklearn.preprocessing import MinMaxScaler\n",
    "import numpy as np\n",
    "import re\n",
    "import json"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "data =  pd.read_csv('../assets/single_extracted_landmarks_inclass_front.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "def landmarks_to_arr(stringarr):\n",
    "    arr = stringarr.replace(\"[\", \"\").replace(\"]\", \"\")\n",
    "    # removing the last bc of persoinal mistake kkkk\n",
    "    return [float(x) for x in arr.split()[:-1]]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "data['landmarks'] = data['landmarks'].apply(landmarks_to_arr)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Preprocessing of the landmarks\n",
    "### Consists on:\n",
    "- Grouping the sequences:\n",
    "   * Shapes must be (9, 10, 51)\n",
    "   * numpy arrays\n",
    "- Isolate each set within the sequence and normalize them.\n",
    "   * Normalize with MinMaxScaler() no params."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(9, 10, 50)\n"
     ]
    }
   ],
   "source": [
    "sequences_list = []\n",
    "SEQUENCE_LENGTH = 10\n",
    "\n",
    "# Grouping by sequences\n",
    "grouped_data = data.groupby(['video', 'group'])\n",
    "for i, group in grouped_data:\n",
    "    landmarks = group['landmarks'].tolist()\n",
    "    if len(landmarks) == 10:\n",
    "        sequences_list.append(landmarks)\n",
    "\n",
    "sequences = np.array(sequences_list)\n",
    "print(sequences.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "scaler = MinMaxScaler()\n",
    "normalized_sequences = np.zeros_like(sequences)\n",
    "# Normalizing\n",
    "for i in range(sequences.shape[0]):\n",
    "    for j in range(sequences.shape[1]):\n",
    "        # Flatten the landmarks for each set within the sequence\n",
    "        landmarks_flattened = np.reshape(sequences[i, j], (-1, 1))\n",
    "        # Normalize the landmarks\n",
    "        landmarks_normalized = scaler.fit_transform(landmarks_flattened)\n",
    "        # Reshape the normalized landmarks back to the original shape\n",
    "        normalized_landmarks = np.reshape(landmarks_normalized, sequences[i, j].shape)\n",
    "        # Update the normalized landmarks in the sequences array\n",
    "        normalized_sequences[i, j] = normalized_landmarks"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(9, 10, 50)"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.array(sequences).shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(1, 10, 50)\n",
      "(1, 10, 50)\n",
      "(1, 10, 50)\n",
      "(1, 10, 50)\n",
      "(1, 10, 50)\n",
      "(1, 10, 50)\n",
      "(1, 10, 50)\n",
      "(1, 10, 50)\n",
      "(1, 10, 50)\n"
     ]
    }
   ],
   "source": [
    "for seq in sequences:\n",
    "    reshaped = seq[np.newaxis, :, :]\n",
    "    print(reshaped.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "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": "3.11.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}