diff --git "a/Apresentacao_do_projeto/.ipynb_checkpoints/Apresenta\303\247\303\243o_Landmarks-checkpoint.ipynb" "b/Apresentacao_do_projeto/.ipynb_checkpoints/Apresenta\303\247\303\243o_Landmarks-checkpoint.ipynb"
index b5f6f3b..139c462 100755
--- "a/Apresentacao_do_projeto/.ipynb_checkpoints/Apresenta\303\247\303\243o_Landmarks-checkpoint.ipynb"
+++ "b/Apresentacao_do_projeto/.ipynb_checkpoints/Apresenta\303\247\303\243o_Landmarks-checkpoint.ipynb"
@@ -77,7 +77,7 @@
     "\n",
     "Representação das distâncias em um rosto desenhado   |  Cálculo das respectivas distâncias\n",
     ":-------------------------------------------------:|:-------------------------:\n",
-    "![face.png](face.png) |  ![dist.png](dist.png)\n",
+    "![ims/face.png](imgs/face.png) |  ![imgs/dist.png](imgs/dist.png)\n",
     "Fonte: [[Feature-Based Face Recognition Using Mixt ure-Distance](https://ieeexplore.ieee.org/document/517076) | Fonte: [Feature-Based Face Recognition Using Mixt ure-Distance](https://ieeexplore.ieee.org/document/517076)\n",
     "\n",
     " Dessa forma foi proposto extrair essas features a partir de uma foto tirada de uma webcam. Algumas considerações importantes, se o modelo não consegue identificar o rosto, as *landmarks* não são expostas, consequentemente o cálculo da distância joga um erro. Como a distância é variável de acordo com a proximidade do rosto à câmera, decidiu-se utilizar uma como normalizadora das outras, no caso, a terceira: distância entre têmporas, por ser constituída majoritáriamente de osso, essa distância é razoavelmente invariante à expressão facial."
@@ -2330,6 +2330,1060 @@
    "source": [
     "std_clf.score(stdtest_x, test_y)"
    ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "# DeepFace - Uma nova abordagem mais tradicional"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    " Os resultados acima são considerávelmente satisfatórios, ainda mais considerando uma abordagem clássica, sem uso de redes neurais. Ainda assim a curiosidade do grupo em relação ao ganho de desempenho possível ao utilizar esse método hoje mais tradicional motivou uma busca por implementações de redes que já foram implementadas e treinadas com tal objetivo, visto que o treino é proibitivo para o grupo, devido à falta de um dataset da magnitude necessária. Assim tentamos nos ater à alguma rede já treinada, tentar modificar sua implementação para nossas necessidades e, se necessário, realizar um ajuste fino de parâmetros para melhorar o desempenho da mesma."
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    " É com esse objetivo que encontramos a interface DeepFace, disponível em: [DeepFace](https://github.com/serengil/deepface). Essa implementação de uma biblioteca que busca utilizar grandes redes de identificação de features faciais, acessar o vetor flatten da rede e realizar a comparação dois a dois destes vetores utilizando alguma métrica de distâncias. Os vetores então podem ser considerados a **codificação das respectivas imagens**, utilizando um threshold definido pela própria biblioteca é possível decidir se as imagens pertencem ou não à mesma pessoa. No caso a biblioteca faz uso de quatro redes diferentes:\n",
+    ">**VGG-Face**: <img src=\"https://www.researchgate.net/profile/Masaki_Nakada/publication/319284653/figure/fig2/AS:655554917720065@1533307880238/VGG-Face-network-architecture.png\" width=100%>\n",
+    "Desenvolvida por Oxford, esta rede tem sua arquitetura baseada na VGG16 e está sob a licença  Creative Commons Attribution, logo não pode ser utilizada em uso comercial.\n",
+    "\n",
+    ">**Facenet**: Rede desenvolvida com base na arquitetura de Zeiler&Fergus, bem como também baseada na inception. Implementada em tensorflow, é uma rede antiga mas tem ótimos resultados. Sua licença é a MIT, portanto pode ser utilizada em uso comercial.\n",
+    "\n",
+    ">**OpenFace**: Implementada em Torch, essa rede é baseada no mesmo paper da Facenet, inclusive foi uma inspiração para a atualização desta pós primeiro lançamento. Está sob a liceça Apache 2.0 e pode ser usada comercialmente.\n",
+    "\n",
+    ">**Deepface**:\n",
+    "<img src = 'https://image.slidesharecdn.com/deepface-170301201504/95/facebook-deep-face-16-638.jpg?cb=1488399466'>\n",
+    "Baseado em um paper de mesmo nome publicado em 2014, esta rede é a implementação da equipe de desenvolvimento do facebook, utilizando a API do Keras. É considerada a mais complexa das quatro aqui citadas e também está sob a licença MIT.\n",
+    "\n",
+    "Assim é possível utilizar qualquer uma das 4 redes citadas acima, e utilizar uma das seguintes métricas: similaridade de cossenos, distância euclidiana, e distância L2. Mas, como temos todos os modelos disponíveis, é possível também realizar um ensemble learning de todos os modelos, utilizando todas as métricas, para fazer a verificação, assim temos o seguinte esquema:\n",
+    "<img src = 'https://raw.githubusercontent.com/serengil/deepface/master/icon/stock-4.jpg'>\n",
+    "\n",
+    "Que é exatamente o que vamos fazer, modificamos a biblioteca original para funcionar melhor com os propósitos do banco, e com propósitos próprios do grupo que vamos mostrar a seguir. A rede ainda possui outras funcionalidades interessantes, como verificação de gênero, idade, etinia, sentimento, etc."
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "## Funcionamento da biblio modificada pelo grupo:"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 1,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "Using TensorFlow backend.\n"
+     ]
+    }
+   ],
+   "source": [
+    "from deepface import DeepFace\n",
+    "import cv2\n",
+    "import numpy as np #para trabalhar com vetores\n",
+    "import glob\n",
+    "import os \n",
+    "import shutil\n",
+    "import pathlib\n",
+    "import os\n",
+    "import pandas as pd"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 2,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "dic = DeepFace.make_dict('/home/fill/Documentos/Turing/VC/DeepFace/Turing_Faces')"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 3,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "{'Rodrigo_Fill.1': '/home/fill/Documentos/Turing/VC/DeepFace/Turing_Faces/Rodrigo_Fill/Fill01.jpeg',\n",
+       " 'Rodrigo_Fill.2': '/home/fill/Documentos/Turing/VC/DeepFace/Turing_Faces/Rodrigo_Fill/Fill02.jpeg',\n",
+       " 'Luisa_Heise.1': '/home/fill/Documentos/Turing/VC/DeepFace/Turing_Faces/Luisa_Heise/Luisa02.jpeg',\n",
+       " 'Luisa_Heise.2': '/home/fill/Documentos/Turing/VC/DeepFace/Turing_Faces/Luisa_Heise/Luisa01.jpeg',\n",
+       " 'Ariel_Guerreiro.1': '/home/fill/Documentos/Turing/VC/DeepFace/Turing_Faces/Ariel_Guerreiro/Ariel01.jpeg',\n",
+       " 'Ariel_Guerreiro.2': '/home/fill/Documentos/Turing/VC/DeepFace/Turing_Faces/Ariel_Guerreiro/Ariel02.jpeg',\n",
+       " 'Guilherme_Goncalves.1': '/home/fill/Documentos/Turing/VC/DeepFace/Turing_Faces/Guilherme_Goncalves/Gui01.jpeg',\n",
+       " 'Guilherme_Goncalves.2': '/home/fill/Documentos/Turing/VC/DeepFace/Turing_Faces/Guilherme_Goncalves/Gui02.jpeg',\n",
+       " 'Camila_Fonseca.1': '/home/fill/Documentos/Turing/VC/DeepFace/Turing_Faces/Camila_Fonseca/Camila02.jpeg',\n",
+       " 'Camila_Fonseca.2': '/home/fill/Documentos/Turing/VC/DeepFace/Turing_Faces/Camila_Fonseca/Camila01.jpeg',\n",
+       " 'Guilherme_Cola.1': '/home/fill/Documentos/Turing/VC/DeepFace/Turing_Faces/Guilherme_Cola/Cola02.jpeg',\n",
+       " 'Guilherme_Cola.2': '/home/fill/Documentos/Turing/VC/DeepFace/Turing_Faces/Guilherme_Cola/Cola01.jpeg',\n",
+       " 'Noel_Viscome.1': '/home/fill/Documentos/Turing/VC/DeepFace/Turing_Faces/Noel_Viscome/Noel02.jpeg',\n",
+       " 'Noel_Viscome.2': '/home/fill/Documentos/Turing/VC/DeepFace/Turing_Faces/Noel_Viscome/Noel01.jpeg',\n",
+       " 'Victor_Jinsi.1': '/home/fill/Documentos/Turing/VC/DeepFace/Turing_Faces/Victor_Jinsi/Victor01.jpeg',\n",
+       " 'Eduardo_Eiras.1': '/home/fill/Documentos/Turing/VC/DeepFace/Turing_Faces/Eduardo_Eiras/Edu06.jpg',\n",
+       " 'Eduardo_Eiras.2': '/home/fill/Documentos/Turing/VC/DeepFace/Turing_Faces/Eduardo_Eiras/Edu05.jpeg',\n",
+       " 'Eduardo_Eiras.3': '/home/fill/Documentos/Turing/VC/DeepFace/Turing_Faces/Eduardo_Eiras/Edu03.jpeg',\n",
+       " 'Eduardo_Eiras.4': '/home/fill/Documentos/Turing/VC/DeepFace/Turing_Faces/Eduardo_Eiras/Edu04.jpeg',\n",
+       " 'Eduardo_Eiras.5': '/home/fill/Documentos/Turing/VC/DeepFace/Turing_Faces/Eduardo_Eiras/Edu01.jpeg',\n",
+       " 'Eduardo_Eiras.6': '/home/fill/Documentos/Turing/VC/DeepFace/Turing_Faces/Eduardo_Eiras/Edu02.jpg',\n",
+       " 'Eduardo_Eiras.7': '/home/fill/Documentos/Turing/VC/DeepFace/Turing_Faces/Eduardo_Eiras/Edu07.jpg',\n",
+       " 'Nelson_Alves.1': '/home/fill/Documentos/Turing/VC/DeepFace/Turing_Faces/Nelson_Alves/Nelson02.jpeg',\n",
+       " 'Nelson_Alves.2': '/home/fill/Documentos/Turing/VC/DeepFace/Turing_Faces/Nelson_Alves/Nelson01.jpeg',\n",
+       " 'Leonardo_Murakami.1': '/home/fill/Documentos/Turing/VC/DeepFace/Turing_Faces/Leonardo_Murakami/Mura01.jpeg',\n",
+       " 'Leonardo_Murakami.2': '/home/fill/Documentos/Turing/VC/DeepFace/Turing_Faces/Leonardo_Murakami/Mura02.jpeg',\n",
+       " 'Felipe_Azank.1': '/home/fill/Documentos/Turing/VC/DeepFace/Turing_Faces/Felipe_Azank/Azank02.jpeg',\n",
+       " 'Felipe_Azank.2': '/home/fill/Documentos/Turing/VC/DeepFace/Turing_Faces/Felipe_Azank/Azank03.jpeg',\n",
+       " 'Felipe_Azank.3': '/home/fill/Documentos/Turing/VC/DeepFace/Turing_Faces/Felipe_Azank/Azank01.jpeg',\n",
+       " 'Enzo_Bustos.1': '/home/fill/Documentos/Turing/VC/DeepFace/Turing_Faces/Enzo_Bustos/Enzo02.jpeg',\n",
+       " 'Enzo_Bustos.2': '/home/fill/Documentos/Turing/VC/DeepFace/Turing_Faces/Enzo_Bustos/Enzo01.jpeg',\n",
+       " 'Paulo_Sestini.1': '/home/fill/Documentos/Turing/VC/DeepFace/Turing_Faces/Paulo_Sestini/Paulo01.jpeg',\n",
+       " 'Paulo_Sestini.2': '/home/fill/Documentos/Turing/VC/DeepFace/Turing_Faces/Paulo_Sestini/Paulo02.jpeg',\n",
+       " 'Alex_Koji.1': '/home/fill/Documentos/Turing/VC/DeepFace/Turing_Faces/Alex_Koji/Koji01.jpeg'}"
+      ]
+     },
+     "execution_count": 3,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "dic"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 4,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Ensemble learning enabled\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "Loading Facebook DeepFace: 100%|██████████| 4/4 [00:43<00:00, 10.87s/it]\n",
+      "Avaliando Hash: 100%|██████████| 34/34 [01:10<00:00,  2.08s/it]\n"
+     ]
+    }
+   ],
+   "source": [
+    "df = DeepFace.save_hash(dic)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 5,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/html": [
+       "<div>\n",
+       "<style scoped>\n",
+       "    .dataframe tbody tr th:only-of-type {\n",
+       "        vertical-align: middle;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe tbody tr th {\n",
+       "        vertical-align: top;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe thead th {\n",
+       "        text-align: right;\n",
+       "    }\n",
+       "</style>\n",
+       "<table border=\"1\" class=\"dataframe\">\n",
+       "  <thead>\n",
+       "    <tr style=\"text-align: right;\">\n",
+       "      <th></th>\n",
+       "      <th>VGG-Face</th>\n",
+       "      <th>Facenet</th>\n",
+       "      <th>OpenFace</th>\n",
+       "      <th>DeepFace</th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>Rodrigo_Fill.1</th>\n",
+       "      <td>[0.011258397, -0.0051986985, -0.0035807253, 0....</td>\n",
+       "      <td>[-0.7107556, 0.2447348, -1.3264471, 0.20392153...</td>\n",
+       "      <td>[0.09385188, -0.07862704, 0.0021203938, 0.0650...</td>\n",
+       "      <td>[0.0, 0.0, 0.66140383, 0.0, 0.6555716, 2.81141...</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>Rodrigo_Fill.2</th>\n",
+       "      <td>[0.01038197, -0.009693551, 0.008820645, 0.0113...</td>\n",
+       "      <td>[0.81892776, 1.5169264, -2.3059661, 0.83081686...</td>\n",
+       "      <td>[0.07494784, -0.013211267, -0.101745866, 0.045...</td>\n",
+       "      <td>[0.0, 0.0, 0.76971376, 1.1680458, 3.7435546, 1...</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>Luisa_Heise.1</th>\n",
+       "      <td>[0.013311128, 0.016905498, 0.017590838, 0.0007...</td>\n",
+       "      <td>[0.15546729, -0.48523352, 1.1752089, -1.828793...</td>\n",
+       "      <td>[0.03517189, 0.043330558, -0.080343574, 0.0138...</td>\n",
+       "      <td>[0.0, 0.0, 0.0, 0.3481037, 1.2885046, 3.498858...</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>Luisa_Heise.2</th>\n",
+       "      <td>[0.0052965735, 0.007177258, -0.0013739086, 0.0...</td>\n",
+       "      <td>[-1.5654796, -1.0087686, 0.77439445, -0.830853...</td>\n",
+       "      <td>[0.015118088, 0.01344703, -0.10854726, -0.0123...</td>\n",
+       "      <td>[0.0, 0.0, 0.0, 0.0, 1.1309611, 4.682887, 2.07...</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>Ariel_Guerreiro.1</th>\n",
+       "      <td>[0.019220604, 0.015955398, 0.025329819, 0.0061...</td>\n",
+       "      <td>[1.8040531, 2.4030004, -1.5023875, 1.736949, 0...</td>\n",
+       "      <td>[-0.03975611, 0.10978596, 0.016175106, 0.07090...</td>\n",
+       "      <td>[0.0, 0.0, 0.0, 0.9227623, 2.601956, 0.0, 0.90...</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>Ariel_Guerreiro.2</th>\n",
+       "      <td>[0.008694698, -0.005788977, -0.0016123804, -0....</td>\n",
+       "      <td>[-0.19725096, 1.6418126, -1.1741674, 0.9594986...</td>\n",
+       "      <td>[0.019117408, 0.18743879, 0.025908452, 0.05127...</td>\n",
+       "      <td>[0.0, 0.0, 0.0, 0.36150473, 0.0, 2.553075, 2.4...</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>Guilherme_Goncalves.1</th>\n",
+       "      <td>[0.0071882443, 6.268373e-05, 0.007376805, 0.00...</td>\n",
+       "      <td>[-0.5617701, 0.15997693, 0.18063924, 0.9567702...</td>\n",
+       "      <td>[-0.041704994, -0.00016754917, 0.035282012, -0...</td>\n",
+       "      <td>[0.0, 0.0, 0.0, 0.03970301, 1.1502867, 0.0, 2....</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>Guilherme_Goncalves.2</th>\n",
+       "      <td>[0.008158567, -0.00033164697, 0.009618739, 0.0...</td>\n",
+       "      <td>[-0.7463429, -0.24570425, 0.20713773, 1.214165...</td>\n",
+       "      <td>[-0.029539527, 0.004947966, 0.0065265903, -0.0...</td>\n",
+       "      <td>[0.0, 0.0, 0.0, 0.0, 0.47723967, 0.0, 3.388195...</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>Camila_Fonseca.1</th>\n",
+       "      <td>[0.020502472, 0.014146947, 0.013597289, -0.001...</td>\n",
+       "      <td>[-1.8097155, -1.8309294, -1.20493, 1.3271358, ...</td>\n",
+       "      <td>[0.051527027, -0.008045665, -0.070917234, 0.00...</td>\n",
+       "      <td>[0.0, 0.0, 0.0, 0.30561933, 3.0658455, 2.70023...</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>Camila_Fonseca.2</th>\n",
+       "      <td>[0.016508777, 0.0141032515, 0.017723288, 0.007...</td>\n",
+       "      <td>[-1.3376508, -1.5952889, 0.25136778, 1.1719253...</td>\n",
+       "      <td>[0.033205245, 0.06364468, -0.036076825, -0.026...</td>\n",
+       "      <td>[0.0, 0.0, 0.0, 0.0, 2.3348985, 5.3893414, 2.8...</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>Guilherme_Cola.1</th>\n",
+       "      <td>[0.011286672, -0.0031716733, 0.006024914, 0.00...</td>\n",
+       "      <td>[-0.10863148, 0.12147777, -2.1797159, 2.553167...</td>\n",
+       "      <td>[0.09351041, 0.023937931, 0.014001634, 0.13893...</td>\n",
+       "      <td>[0.0, 0.0, 0.0, 0.0, 0.0, 3.8363554, 0.0, 0.0,...</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>Guilherme_Cola.2</th>\n",
+       "      <td>[0.016229464, -0.006794181, 0.0052082087, -1.6...</td>\n",
+       "      <td>[0.19653642, -0.065777265, -2.1988037, 2.93330...</td>\n",
+       "      <td>[0.07559654, 0.0140757235, -0.089574635, 0.076...</td>\n",
+       "      <td>[0.0, 0.0, 0.0, 0.0, 0.8358648, 6.13692, 1.777...</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>Noel_Viscome.1</th>\n",
+       "      <td>[0.007722295, -0.0032768892, 0.01261278, 0.015...</td>\n",
+       "      <td>[-1.7278033, 0.26465106, 0.6038073, 0.36027545...</td>\n",
+       "      <td>[0.038704526, 0.07394758, -0.09165, 0.04594578...</td>\n",
+       "      <td>[0.0, 0.0, 0.0, 0.064310044, 1.9042714, 2.0811...</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>Noel_Viscome.2</th>\n",
+       "      <td>[0.009547951, 0.0034925663, 0.01011407, 0.0061...</td>\n",
+       "      <td>[-0.5509963, 0.74306023, 0.14488125, -0.936631...</td>\n",
+       "      <td>[0.051902585, -0.023795338, -0.08508713, -0.03...</td>\n",
+       "      <td>[0.0, 0.0, 0.0, 0.0, 3.124282, 2.3894014, 3.08...</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>Victor_Jinsi.1</th>\n",
+       "      <td>[0.020444352, 0.010158352, 0.00625316, 0.01316...</td>\n",
+       "      <td>[0.5734688, 0.86270785, -0.64202434, 1.1462065...</td>\n",
+       "      <td>[-0.01967564, 0.0672678, 0.011682327, 0.033551...</td>\n",
+       "      <td>[0.0, 0.0, 0.0, 0.0, 2.0280519, 0.0, 1.4935838...</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>Eduardo_Eiras.1</th>\n",
+       "      <td>[0.021600582, 0.006747674, 0.013444074, 0.0118...</td>\n",
+       "      <td>[-0.34935245, -0.37163797, -0.39313334, 0.7243...</td>\n",
+       "      <td>[-0.005565289, 0.051990893, -0.041520093, -0.0...</td>\n",
+       "      <td>[0.0, 0.0, 0.0, 0.26732165, 1.0506326, 0.35762...</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>Eduardo_Eiras.2</th>\n",
+       "      <td>[0.027794681, 0.012595378, 0.015973512, 0.0243...</td>\n",
+       "      <td>[0.4253676, 0.68349314, 0.5815365, -0.06460958...</td>\n",
+       "      <td>[-0.0114504425, -0.02118617, -0.09935128, -0.0...</td>\n",
+       "      <td>[0.0, 0.0, 0.0, 0.0, 3.7076707, 0.6628028, 2.7...</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>Eduardo_Eiras.3</th>\n",
+       "      <td>[0.023938581, -0.005019143, -0.0041286526, 0.0...</td>\n",
+       "      <td>[-1.293877, -0.1234024, -1.4063007, 1.3587369,...</td>\n",
+       "      <td>[-0.02248524, 0.14344844, 0.04551649, 0.008724...</td>\n",
+       "      <td>[0.0, 0.0, 0.0, 0.0, 2.8374405, 2.356803, 0.70...</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>Eduardo_Eiras.5</th>\n",
+       "      <td>[0.014218462, 0.0028408384, 0.012237807, 0.018...</td>\n",
+       "      <td>[-1.3807191, 0.006177306, -0.50413, 1.6096864,...</td>\n",
+       "      <td>[-0.004796549, 0.002740741, -0.01581459, -0.02...</td>\n",
+       "      <td>[0.0, 0.0, 0.0, 0.0, 3.0053782, 0.65704703, 2....</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>Eduardo_Eiras.6</th>\n",
+       "      <td>[0.023989934, 0.014054777, 0.004245261, 0.0108...</td>\n",
+       "      <td>[-0.7857877, 0.35950303, -0.56709725, 0.465594...</td>\n",
+       "      <td>[-0.03314291, -0.03338445, -0.0071588843, -0.0...</td>\n",
+       "      <td>[0.0, 0.0, 0.0, 0.27838233, 0.0, 1.7221168, 1....</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>Eduardo_Eiras.7</th>\n",
+       "      <td>[0.023237122, 0.011033161, 0.016982649, 0.0112...</td>\n",
+       "      <td>[-0.9716029, -0.26354784, -0.4264811, 1.663771...</td>\n",
+       "      <td>[-0.06154397, 0.031329878, 0.051613558, -0.094...</td>\n",
+       "      <td>[0.0, 0.0, 0.0, 0.334877, 0.69340277, 0.0, 0.0...</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>Nelson_Alves.1</th>\n",
+       "      <td>[0.011202453, -0.011317662, 0.006242816, -0.00...</td>\n",
+       "      <td>[0.2442035, -0.018297125, -1.086241, 1.2000073...</td>\n",
+       "      <td>[0.0517047, 0.13678262, 0.06404092, 0.07557396...</td>\n",
+       "      <td>[0.0, 0.0, 0.0, 0.0, 1.1573782, 2.4152763, 1.9...</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>Leonardo_Murakami.1</th>\n",
+       "      <td>[0.017816652, -0.0014210419, 0.0012131145, 0.0...</td>\n",
+       "      <td>[-1.7009012, -0.6807573, 0.14056861, -0.153483...</td>\n",
+       "      <td>[-0.015473774, -0.0020993, -0.03174869, -0.055...</td>\n",
+       "      <td>[0.0, 0.0, 0.0, 0.62446773, 0.0, 0.0, 1.890828...</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>Leonardo_Murakami.2</th>\n",
+       "      <td>[0.01864994, 0.0052387556, -0.0028785854, 0.01...</td>\n",
+       "      <td>[-1.7920934, -1.4230772, 0.6691602, -0.2342174...</td>\n",
+       "      <td>[0.035412, 0.00017503757, -0.043976996, -0.071...</td>\n",
+       "      <td>[0.0, 0.0, 0.0, 0.6731644, 0.0, 0.0, 0.6979854...</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>Enzo_Bustos.1</th>\n",
+       "      <td>[0.02399399, 0.0052210162, -0.015569969, 0.012...</td>\n",
+       "      <td>[-2.1904962, -1.3387287, 0.13840455, 0.6652469...</td>\n",
+       "      <td>[0.10071324, -0.036814034, -0.1275556, 0.04338...</td>\n",
+       "      <td>[0.0, 0.0, 0.0, 0.0, 4.0648704, 8.178255, 1.45...</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>Enzo_Bustos.2</th>\n",
+       "      <td>[0.019614054, 0.0023429843, -0.0011203118, 0.0...</td>\n",
+       "      <td>[0.29559124, -0.25272578, 0.33746648, 1.579556...</td>\n",
+       "      <td>[0.13314737, 0.06992418, -0.15777715, 0.004923...</td>\n",
+       "      <td>[0.0, 0.0, 0.0, 0.0, 2.9943137, 5.8091803, 3.4...</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>Paulo_Sestini.1</th>\n",
+       "      <td>[0.013956297, 0.00030987675, 0.0015853783, 0.0...</td>\n",
+       "      <td>[-1.698124, -0.39025182, -1.4879581, 1.7800299...</td>\n",
+       "      <td>[-0.0883103, -0.007851686, 0.0411439, -0.07708...</td>\n",
+       "      <td>[0.0, 0.0, 0.0, 0.0, 1.3161561, 0.009125233, 1...</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>Paulo_Sestini.2</th>\n",
+       "      <td>[0.015406041, -0.00386795, -0.0001494264, 0.00...</td>\n",
+       "      <td>[-0.886387, 0.30212864, 0.016729653, 2.1621618...</td>\n",
+       "      <td>[-0.043899756, 0.054894928, 0.010215374, -0.05...</td>\n",
+       "      <td>[0.0, 0.0, 0.0, 0.0, 1.3580369, 0.0, 2.9301717...</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>Alex_Koji.1</th>\n",
+       "      <td>[0.013313855, 0.016293736, -0.0011014326, 0.00...</td>\n",
+       "      <td>[0.06953996, -1.1348931, -1.6284283, 0.6346901...</td>\n",
+       "      <td>[-0.004252907, 0.17901103, 0.0792524, 0.141097...</td>\n",
+       "      <td>[0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 1.8053236, 0.0,...</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "                                                                VGG-Face  \\\n",
+       "Rodrigo_Fill.1         [0.011258397, -0.0051986985, -0.0035807253, 0....   \n",
+       "Rodrigo_Fill.2         [0.01038197, -0.009693551, 0.008820645, 0.0113...   \n",
+       "Luisa_Heise.1          [0.013311128, 0.016905498, 0.017590838, 0.0007...   \n",
+       "Luisa_Heise.2          [0.0052965735, 0.007177258, -0.0013739086, 0.0...   \n",
+       "Ariel_Guerreiro.1      [0.019220604, 0.015955398, 0.025329819, 0.0061...   \n",
+       "Ariel_Guerreiro.2      [0.008694698, -0.005788977, -0.0016123804, -0....   \n",
+       "Guilherme_Goncalves.1  [0.0071882443, 6.268373e-05, 0.007376805, 0.00...   \n",
+       "Guilherme_Goncalves.2  [0.008158567, -0.00033164697, 0.009618739, 0.0...   \n",
+       "Camila_Fonseca.1       [0.020502472, 0.014146947, 0.013597289, -0.001...   \n",
+       "Camila_Fonseca.2       [0.016508777, 0.0141032515, 0.017723288, 0.007...   \n",
+       "Guilherme_Cola.1       [0.011286672, -0.0031716733, 0.006024914, 0.00...   \n",
+       "Guilherme_Cola.2       [0.016229464, -0.006794181, 0.0052082087, -1.6...   \n",
+       "Noel_Viscome.1         [0.007722295, -0.0032768892, 0.01261278, 0.015...   \n",
+       "Noel_Viscome.2         [0.009547951, 0.0034925663, 0.01011407, 0.0061...   \n",
+       "Victor_Jinsi.1         [0.020444352, 0.010158352, 0.00625316, 0.01316...   \n",
+       "Eduardo_Eiras.1        [0.021600582, 0.006747674, 0.013444074, 0.0118...   \n",
+       "Eduardo_Eiras.2        [0.027794681, 0.012595378, 0.015973512, 0.0243...   \n",
+       "Eduardo_Eiras.3        [0.023938581, -0.005019143, -0.0041286526, 0.0...   \n",
+       "Eduardo_Eiras.5        [0.014218462, 0.0028408384, 0.012237807, 0.018...   \n",
+       "Eduardo_Eiras.6        [0.023989934, 0.014054777, 0.004245261, 0.0108...   \n",
+       "Eduardo_Eiras.7        [0.023237122, 0.011033161, 0.016982649, 0.0112...   \n",
+       "Nelson_Alves.1         [0.011202453, -0.011317662, 0.006242816, -0.00...   \n",
+       "Leonardo_Murakami.1    [0.017816652, -0.0014210419, 0.0012131145, 0.0...   \n",
+       "Leonardo_Murakami.2    [0.01864994, 0.0052387556, -0.0028785854, 0.01...   \n",
+       "Enzo_Bustos.1          [0.02399399, 0.0052210162, -0.015569969, 0.012...   \n",
+       "Enzo_Bustos.2          [0.019614054, 0.0023429843, -0.0011203118, 0.0...   \n",
+       "Paulo_Sestini.1        [0.013956297, 0.00030987675, 0.0015853783, 0.0...   \n",
+       "Paulo_Sestini.2        [0.015406041, -0.00386795, -0.0001494264, 0.00...   \n",
+       "Alex_Koji.1            [0.013313855, 0.016293736, -0.0011014326, 0.00...   \n",
+       "\n",
+       "                                                                 Facenet  \\\n",
+       "Rodrigo_Fill.1         [-0.7107556, 0.2447348, -1.3264471, 0.20392153...   \n",
+       "Rodrigo_Fill.2         [0.81892776, 1.5169264, -2.3059661, 0.83081686...   \n",
+       "Luisa_Heise.1          [0.15546729, -0.48523352, 1.1752089, -1.828793...   \n",
+       "Luisa_Heise.2          [-1.5654796, -1.0087686, 0.77439445, -0.830853...   \n",
+       "Ariel_Guerreiro.1      [1.8040531, 2.4030004, -1.5023875, 1.736949, 0...   \n",
+       "Ariel_Guerreiro.2      [-0.19725096, 1.6418126, -1.1741674, 0.9594986...   \n",
+       "Guilherme_Goncalves.1  [-0.5617701, 0.15997693, 0.18063924, 0.9567702...   \n",
+       "Guilherme_Goncalves.2  [-0.7463429, -0.24570425, 0.20713773, 1.214165...   \n",
+       "Camila_Fonseca.1       [-1.8097155, -1.8309294, -1.20493, 1.3271358, ...   \n",
+       "Camila_Fonseca.2       [-1.3376508, -1.5952889, 0.25136778, 1.1719253...   \n",
+       "Guilherme_Cola.1       [-0.10863148, 0.12147777, -2.1797159, 2.553167...   \n",
+       "Guilherme_Cola.2       [0.19653642, -0.065777265, -2.1988037, 2.93330...   \n",
+       "Noel_Viscome.1         [-1.7278033, 0.26465106, 0.6038073, 0.36027545...   \n",
+       "Noel_Viscome.2         [-0.5509963, 0.74306023, 0.14488125, -0.936631...   \n",
+       "Victor_Jinsi.1         [0.5734688, 0.86270785, -0.64202434, 1.1462065...   \n",
+       "Eduardo_Eiras.1        [-0.34935245, -0.37163797, -0.39313334, 0.7243...   \n",
+       "Eduardo_Eiras.2        [0.4253676, 0.68349314, 0.5815365, -0.06460958...   \n",
+       "Eduardo_Eiras.3        [-1.293877, -0.1234024, -1.4063007, 1.3587369,...   \n",
+       "Eduardo_Eiras.5        [-1.3807191, 0.006177306, -0.50413, 1.6096864,...   \n",
+       "Eduardo_Eiras.6        [-0.7857877, 0.35950303, -0.56709725, 0.465594...   \n",
+       "Eduardo_Eiras.7        [-0.9716029, -0.26354784, -0.4264811, 1.663771...   \n",
+       "Nelson_Alves.1         [0.2442035, -0.018297125, -1.086241, 1.2000073...   \n",
+       "Leonardo_Murakami.1    [-1.7009012, -0.6807573, 0.14056861, -0.153483...   \n",
+       "Leonardo_Murakami.2    [-1.7920934, -1.4230772, 0.6691602, -0.2342174...   \n",
+       "Enzo_Bustos.1          [-2.1904962, -1.3387287, 0.13840455, 0.6652469...   \n",
+       "Enzo_Bustos.2          [0.29559124, -0.25272578, 0.33746648, 1.579556...   \n",
+       "Paulo_Sestini.1        [-1.698124, -0.39025182, -1.4879581, 1.7800299...   \n",
+       "Paulo_Sestini.2        [-0.886387, 0.30212864, 0.016729653, 2.1621618...   \n",
+       "Alex_Koji.1            [0.06953996, -1.1348931, -1.6284283, 0.6346901...   \n",
+       "\n",
+       "                                                                OpenFace  \\\n",
+       "Rodrigo_Fill.1         [0.09385188, -0.07862704, 0.0021203938, 0.0650...   \n",
+       "Rodrigo_Fill.2         [0.07494784, -0.013211267, -0.101745866, 0.045...   \n",
+       "Luisa_Heise.1          [0.03517189, 0.043330558, -0.080343574, 0.0138...   \n",
+       "Luisa_Heise.2          [0.015118088, 0.01344703, -0.10854726, -0.0123...   \n",
+       "Ariel_Guerreiro.1      [-0.03975611, 0.10978596, 0.016175106, 0.07090...   \n",
+       "Ariel_Guerreiro.2      [0.019117408, 0.18743879, 0.025908452, 0.05127...   \n",
+       "Guilherme_Goncalves.1  [-0.041704994, -0.00016754917, 0.035282012, -0...   \n",
+       "Guilherme_Goncalves.2  [-0.029539527, 0.004947966, 0.0065265903, -0.0...   \n",
+       "Camila_Fonseca.1       [0.051527027, -0.008045665, -0.070917234, 0.00...   \n",
+       "Camila_Fonseca.2       [0.033205245, 0.06364468, -0.036076825, -0.026...   \n",
+       "Guilherme_Cola.1       [0.09351041, 0.023937931, 0.014001634, 0.13893...   \n",
+       "Guilherme_Cola.2       [0.07559654, 0.0140757235, -0.089574635, 0.076...   \n",
+       "Noel_Viscome.1         [0.038704526, 0.07394758, -0.09165, 0.04594578...   \n",
+       "Noel_Viscome.2         [0.051902585, -0.023795338, -0.08508713, -0.03...   \n",
+       "Victor_Jinsi.1         [-0.01967564, 0.0672678, 0.011682327, 0.033551...   \n",
+       "Eduardo_Eiras.1        [-0.005565289, 0.051990893, -0.041520093, -0.0...   \n",
+       "Eduardo_Eiras.2        [-0.0114504425, -0.02118617, -0.09935128, -0.0...   \n",
+       "Eduardo_Eiras.3        [-0.02248524, 0.14344844, 0.04551649, 0.008724...   \n",
+       "Eduardo_Eiras.5        [-0.004796549, 0.002740741, -0.01581459, -0.02...   \n",
+       "Eduardo_Eiras.6        [-0.03314291, -0.03338445, -0.0071588843, -0.0...   \n",
+       "Eduardo_Eiras.7        [-0.06154397, 0.031329878, 0.051613558, -0.094...   \n",
+       "Nelson_Alves.1         [0.0517047, 0.13678262, 0.06404092, 0.07557396...   \n",
+       "Leonardo_Murakami.1    [-0.015473774, -0.0020993, -0.03174869, -0.055...   \n",
+       "Leonardo_Murakami.2    [0.035412, 0.00017503757, -0.043976996, -0.071...   \n",
+       "Enzo_Bustos.1          [0.10071324, -0.036814034, -0.1275556, 0.04338...   \n",
+       "Enzo_Bustos.2          [0.13314737, 0.06992418, -0.15777715, 0.004923...   \n",
+       "Paulo_Sestini.1        [-0.0883103, -0.007851686, 0.0411439, -0.07708...   \n",
+       "Paulo_Sestini.2        [-0.043899756, 0.054894928, 0.010215374, -0.05...   \n",
+       "Alex_Koji.1            [-0.004252907, 0.17901103, 0.0792524, 0.141097...   \n",
+       "\n",
+       "                                                                DeepFace  \n",
+       "Rodrigo_Fill.1         [0.0, 0.0, 0.66140383, 0.0, 0.6555716, 2.81141...  \n",
+       "Rodrigo_Fill.2         [0.0, 0.0, 0.76971376, 1.1680458, 3.7435546, 1...  \n",
+       "Luisa_Heise.1          [0.0, 0.0, 0.0, 0.3481037, 1.2885046, 3.498858...  \n",
+       "Luisa_Heise.2          [0.0, 0.0, 0.0, 0.0, 1.1309611, 4.682887, 2.07...  \n",
+       "Ariel_Guerreiro.1      [0.0, 0.0, 0.0, 0.9227623, 2.601956, 0.0, 0.90...  \n",
+       "Ariel_Guerreiro.2      [0.0, 0.0, 0.0, 0.36150473, 0.0, 2.553075, 2.4...  \n",
+       "Guilherme_Goncalves.1  [0.0, 0.0, 0.0, 0.03970301, 1.1502867, 0.0, 2....  \n",
+       "Guilherme_Goncalves.2  [0.0, 0.0, 0.0, 0.0, 0.47723967, 0.0, 3.388195...  \n",
+       "Camila_Fonseca.1       [0.0, 0.0, 0.0, 0.30561933, 3.0658455, 2.70023...  \n",
+       "Camila_Fonseca.2       [0.0, 0.0, 0.0, 0.0, 2.3348985, 5.3893414, 2.8...  \n",
+       "Guilherme_Cola.1       [0.0, 0.0, 0.0, 0.0, 0.0, 3.8363554, 0.0, 0.0,...  \n",
+       "Guilherme_Cola.2       [0.0, 0.0, 0.0, 0.0, 0.8358648, 6.13692, 1.777...  \n",
+       "Noel_Viscome.1         [0.0, 0.0, 0.0, 0.064310044, 1.9042714, 2.0811...  \n",
+       "Noel_Viscome.2         [0.0, 0.0, 0.0, 0.0, 3.124282, 2.3894014, 3.08...  \n",
+       "Victor_Jinsi.1         [0.0, 0.0, 0.0, 0.0, 2.0280519, 0.0, 1.4935838...  \n",
+       "Eduardo_Eiras.1        [0.0, 0.0, 0.0, 0.26732165, 1.0506326, 0.35762...  \n",
+       "Eduardo_Eiras.2        [0.0, 0.0, 0.0, 0.0, 3.7076707, 0.6628028, 2.7...  \n",
+       "Eduardo_Eiras.3        [0.0, 0.0, 0.0, 0.0, 2.8374405, 2.356803, 0.70...  \n",
+       "Eduardo_Eiras.5        [0.0, 0.0, 0.0, 0.0, 3.0053782, 0.65704703, 2....  \n",
+       "Eduardo_Eiras.6        [0.0, 0.0, 0.0, 0.27838233, 0.0, 1.7221168, 1....  \n",
+       "Eduardo_Eiras.7        [0.0, 0.0, 0.0, 0.334877, 0.69340277, 0.0, 0.0...  \n",
+       "Nelson_Alves.1         [0.0, 0.0, 0.0, 0.0, 1.1573782, 2.4152763, 1.9...  \n",
+       "Leonardo_Murakami.1    [0.0, 0.0, 0.0, 0.62446773, 0.0, 0.0, 1.890828...  \n",
+       "Leonardo_Murakami.2    [0.0, 0.0, 0.0, 0.6731644, 0.0, 0.0, 0.6979854...  \n",
+       "Enzo_Bustos.1          [0.0, 0.0, 0.0, 0.0, 4.0648704, 8.178255, 1.45...  \n",
+       "Enzo_Bustos.2          [0.0, 0.0, 0.0, 0.0, 2.9943137, 5.8091803, 3.4...  \n",
+       "Paulo_Sestini.1        [0.0, 0.0, 0.0, 0.0, 1.3161561, 0.009125233, 1...  \n",
+       "Paulo_Sestini.2        [0.0, 0.0, 0.0, 0.0, 1.3580369, 0.0, 2.9301717...  \n",
+       "Alex_Koji.1            [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 1.8053236, 0.0,...  "
+      ]
+     },
+     "execution_count": 5,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "df"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 6,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "Avaliando cada pessoa em relação à todas as outras: 100%|██████████| 29/29 [00:00<00:00, 99.78it/s]\n"
+     ]
+    }
+   ],
+   "source": [
+    "metrics_df = DeepFace.metrics_dataframe(df, one_compair = False)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 7,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/html": [
+       "<div>\n",
+       "<style scoped>\n",
+       "    .dataframe tbody tr th:only-of-type {\n",
+       "        vertical-align: middle;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe tbody tr th {\n",
+       "        vertical-align: top;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe thead th {\n",
+       "        text-align: right;\n",
+       "    }\n",
+       "</style>\n",
+       "<table border=\"1\" class=\"dataframe\">\n",
+       "  <thead>\n",
+       "    <tr style=\"text-align: right;\">\n",
+       "      <th></th>\n",
+       "      <th>Cosine with VGG-Face</th>\n",
+       "      <th>Euclidean with VGG-Face</th>\n",
+       "      <th>L2 with VGG-Face</th>\n",
+       "      <th>Cosine with Facenet</th>\n",
+       "      <th>Euclidean with Facenet</th>\n",
+       "      <th>L2 with Facenet</th>\n",
+       "      <th>Cosine with OpenFace</th>\n",
+       "      <th>Euclidean with OpenFace</th>\n",
+       "      <th>L2 with OpenFace</th>\n",
+       "      <th>Cosine with DeepFace</th>\n",
+       "      <th>Euclidean with DeepFace</th>\n",
+       "      <th>L2 with DeepFace</th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>Rodrigo_Fill.1-Rodrigo_Fill.1</th>\n",
+       "      <td>5.96046e-08</td>\n",
+       "      <td>0</td>\n",
+       "      <td>0</td>\n",
+       "      <td>0</td>\n",
+       "      <td>0</td>\n",
+       "      <td>0</td>\n",
+       "      <td>5.96046e-08</td>\n",
+       "      <td>0</td>\n",
+       "      <td>0</td>\n",
+       "      <td>-1.19209e-07</td>\n",
+       "      <td>0</td>\n",
+       "      <td>0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>Rodrigo_Fill.1-Rodrigo_Fill.2</th>\n",
+       "      <td>0.338685</td>\n",
+       "      <td>0.308737</td>\n",
+       "      <td>0.823025</td>\n",
+       "      <td>0.335995</td>\n",
+       "      <td>10.1857</td>\n",
+       "      <td>0.81975</td>\n",
+       "      <td>0.165649</td>\n",
+       "      <td>0.575585</td>\n",
+       "      <td>0.575585</td>\n",
+       "      <td>0.205131</td>\n",
+       "      <td>57.1585</td>\n",
+       "      <td>0.640517</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>Rodrigo_Fill.1-Luisa_Heise.1</th>\n",
+       "      <td>0.657116</td>\n",
+       "      <td>0.718266</td>\n",
+       "      <td>1.1464</td>\n",
+       "      <td>0.771055</td>\n",
+       "      <td>15.2804</td>\n",
+       "      <td>1.24182</td>\n",
+       "      <td>0.366857</td>\n",
+       "      <td>0.856571</td>\n",
+       "      <td>0.856571</td>\n",
+       "      <td>0.400876</td>\n",
+       "      <td>68.9361</td>\n",
+       "      <td>0.895406</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>Rodrigo_Fill.1-Luisa_Heise.2</th>\n",
+       "      <td>0.605889</td>\n",
+       "      <td>0.586127</td>\n",
+       "      <td>1.10081</td>\n",
+       "      <td>0.778835</td>\n",
+       "      <td>14.8838</td>\n",
+       "      <td>1.24807</td>\n",
+       "      <td>0.613198</td>\n",
+       "      <td>1.10743</td>\n",
+       "      <td>1.10743</td>\n",
+       "      <td>0.457767</td>\n",
+       "      <td>73.5529</td>\n",
+       "      <td>0.956835</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>Rodrigo_Fill.1-Ariel_Guerreiro.1</th>\n",
+       "      <td>0.594689</td>\n",
+       "      <td>0.725428</td>\n",
+       "      <td>1.09059</td>\n",
+       "      <td>1.01108</td>\n",
+       "      <td>17.6072</td>\n",
+       "      <td>1.42203</td>\n",
+       "      <td>0.655014</td>\n",
+       "      <td>1.14456</td>\n",
+       "      <td>1.14456</td>\n",
+       "      <td>0.344836</td>\n",
+       "      <td>60.5003</td>\n",
+       "      <td>0.830464</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>...</th>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>Alex_Koji.1-Enzo_Bustos.1</th>\n",
+       "      <td>0.550818</td>\n",
+       "      <td>0.76577</td>\n",
+       "      <td>1.04959</td>\n",
+       "      <td>0.875381</td>\n",
+       "      <td>16.0916</td>\n",
+       "      <td>1.32316</td>\n",
+       "      <td>0.579297</td>\n",
+       "      <td>1.07638</td>\n",
+       "      <td>1.07638</td>\n",
+       "      <td>0.502343</td>\n",
+       "      <td>86.6502</td>\n",
+       "      <td>1.00234</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>Alex_Koji.1-Enzo_Bustos.2</th>\n",
+       "      <td>0.510256</td>\n",
+       "      <td>0.748047</td>\n",
+       "      <td>1.0102</td>\n",
+       "      <td>0.694072</td>\n",
+       "      <td>14.8561</td>\n",
+       "      <td>1.1782</td>\n",
+       "      <td>0.525571</td>\n",
+       "      <td>1.02525</td>\n",
+       "      <td>1.02525</td>\n",
+       "      <td>0.425999</td>\n",
+       "      <td>86.5941</td>\n",
+       "      <td>0.923037</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>Alex_Koji.1-Paulo_Sestini.1</th>\n",
+       "      <td>0.509008</td>\n",
+       "      <td>0.679595</td>\n",
+       "      <td>1.00897</td>\n",
+       "      <td>0.701904</td>\n",
+       "      <td>14.3709</td>\n",
+       "      <td>1.18482</td>\n",
+       "      <td>0.638041</td>\n",
+       "      <td>1.12964</td>\n",
+       "      <td>1.12964</td>\n",
+       "      <td>0.415094</td>\n",
+       "      <td>69.9926</td>\n",
+       "      <td>0.911146</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>Alex_Koji.1-Paulo_Sestini.2</th>\n",
+       "      <td>0.462776</td>\n",
+       "      <td>0.738674</td>\n",
+       "      <td>0.962056</td>\n",
+       "      <td>0.631634</td>\n",
+       "      <td>13.7257</td>\n",
+       "      <td>1.12395</td>\n",
+       "      <td>0.497972</td>\n",
+       "      <td>0.99797</td>\n",
+       "      <td>0.99797</td>\n",
+       "      <td>0.312999</td>\n",
+       "      <td>63.095</td>\n",
+       "      <td>0.7912</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>Alex_Koji.1-Alex_Koji.1</th>\n",
+       "      <td>0</td>\n",
+       "      <td>0</td>\n",
+       "      <td>0</td>\n",
+       "      <td>-1.19209e-07</td>\n",
+       "      <td>0</td>\n",
+       "      <td>0</td>\n",
+       "      <td>-1.19209e-07</td>\n",
+       "      <td>0</td>\n",
+       "      <td>0</td>\n",
+       "      <td>5.96046e-08</td>\n",
+       "      <td>0</td>\n",
+       "      <td>0</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "<p>841 rows × 12 columns</p>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "                                 Cosine with VGG-Face Euclidean with VGG-Face  \\\n",
+       "Rodrigo_Fill.1-Rodrigo_Fill.1             5.96046e-08                       0   \n",
+       "Rodrigo_Fill.1-Rodrigo_Fill.2                0.338685                0.308737   \n",
+       "Rodrigo_Fill.1-Luisa_Heise.1                 0.657116                0.718266   \n",
+       "Rodrigo_Fill.1-Luisa_Heise.2                 0.605889                0.586127   \n",
+       "Rodrigo_Fill.1-Ariel_Guerreiro.1             0.594689                0.725428   \n",
+       "...                                               ...                     ...   \n",
+       "Alex_Koji.1-Enzo_Bustos.1                    0.550818                 0.76577   \n",
+       "Alex_Koji.1-Enzo_Bustos.2                    0.510256                0.748047   \n",
+       "Alex_Koji.1-Paulo_Sestini.1                  0.509008                0.679595   \n",
+       "Alex_Koji.1-Paulo_Sestini.2                  0.462776                0.738674   \n",
+       "Alex_Koji.1-Alex_Koji.1                             0                       0   \n",
+       "\n",
+       "                                 L2 with VGG-Face Cosine with Facenet  \\\n",
+       "Rodrigo_Fill.1-Rodrigo_Fill.1                   0                   0   \n",
+       "Rodrigo_Fill.1-Rodrigo_Fill.2            0.823025            0.335995   \n",
+       "Rodrigo_Fill.1-Luisa_Heise.1               1.1464            0.771055   \n",
+       "Rodrigo_Fill.1-Luisa_Heise.2              1.10081            0.778835   \n",
+       "Rodrigo_Fill.1-Ariel_Guerreiro.1          1.09059             1.01108   \n",
+       "...                                           ...                 ...   \n",
+       "Alex_Koji.1-Enzo_Bustos.1                 1.04959            0.875381   \n",
+       "Alex_Koji.1-Enzo_Bustos.2                  1.0102            0.694072   \n",
+       "Alex_Koji.1-Paulo_Sestini.1               1.00897            0.701904   \n",
+       "Alex_Koji.1-Paulo_Sestini.2              0.962056            0.631634   \n",
+       "Alex_Koji.1-Alex_Koji.1                         0        -1.19209e-07   \n",
+       "\n",
+       "                                 Euclidean with Facenet L2 with Facenet  \\\n",
+       "Rodrigo_Fill.1-Rodrigo_Fill.1                         0               0   \n",
+       "Rodrigo_Fill.1-Rodrigo_Fill.2                   10.1857         0.81975   \n",
+       "Rodrigo_Fill.1-Luisa_Heise.1                    15.2804         1.24182   \n",
+       "Rodrigo_Fill.1-Luisa_Heise.2                    14.8838         1.24807   \n",
+       "Rodrigo_Fill.1-Ariel_Guerreiro.1                17.6072         1.42203   \n",
+       "...                                                 ...             ...   \n",
+       "Alex_Koji.1-Enzo_Bustos.1                       16.0916         1.32316   \n",
+       "Alex_Koji.1-Enzo_Bustos.2                       14.8561          1.1782   \n",
+       "Alex_Koji.1-Paulo_Sestini.1                     14.3709         1.18482   \n",
+       "Alex_Koji.1-Paulo_Sestini.2                     13.7257         1.12395   \n",
+       "Alex_Koji.1-Alex_Koji.1                               0               0   \n",
+       "\n",
+       "                                 Cosine with OpenFace Euclidean with OpenFace  \\\n",
+       "Rodrigo_Fill.1-Rodrigo_Fill.1             5.96046e-08                       0   \n",
+       "Rodrigo_Fill.1-Rodrigo_Fill.2                0.165649                0.575585   \n",
+       "Rodrigo_Fill.1-Luisa_Heise.1                 0.366857                0.856571   \n",
+       "Rodrigo_Fill.1-Luisa_Heise.2                 0.613198                 1.10743   \n",
+       "Rodrigo_Fill.1-Ariel_Guerreiro.1             0.655014                 1.14456   \n",
+       "...                                               ...                     ...   \n",
+       "Alex_Koji.1-Enzo_Bustos.1                    0.579297                 1.07638   \n",
+       "Alex_Koji.1-Enzo_Bustos.2                    0.525571                 1.02525   \n",
+       "Alex_Koji.1-Paulo_Sestini.1                  0.638041                 1.12964   \n",
+       "Alex_Koji.1-Paulo_Sestini.2                  0.497972                 0.99797   \n",
+       "Alex_Koji.1-Alex_Koji.1                  -1.19209e-07                       0   \n",
+       "\n",
+       "                                 L2 with OpenFace Cosine with DeepFace  \\\n",
+       "Rodrigo_Fill.1-Rodrigo_Fill.1                   0         -1.19209e-07   \n",
+       "Rodrigo_Fill.1-Rodrigo_Fill.2            0.575585             0.205131   \n",
+       "Rodrigo_Fill.1-Luisa_Heise.1             0.856571             0.400876   \n",
+       "Rodrigo_Fill.1-Luisa_Heise.2              1.10743             0.457767   \n",
+       "Rodrigo_Fill.1-Ariel_Guerreiro.1          1.14456             0.344836   \n",
+       "...                                           ...                  ...   \n",
+       "Alex_Koji.1-Enzo_Bustos.1                 1.07638             0.502343   \n",
+       "Alex_Koji.1-Enzo_Bustos.2                 1.02525             0.425999   \n",
+       "Alex_Koji.1-Paulo_Sestini.1               1.12964             0.415094   \n",
+       "Alex_Koji.1-Paulo_Sestini.2               0.99797             0.312999   \n",
+       "Alex_Koji.1-Alex_Koji.1                         0          5.96046e-08   \n",
+       "\n",
+       "                                 Euclidean with DeepFace L2 with DeepFace  \n",
+       "Rodrigo_Fill.1-Rodrigo_Fill.1                          0                0  \n",
+       "Rodrigo_Fill.1-Rodrigo_Fill.2                    57.1585         0.640517  \n",
+       "Rodrigo_Fill.1-Luisa_Heise.1                     68.9361         0.895406  \n",
+       "Rodrigo_Fill.1-Luisa_Heise.2                     73.5529         0.956835  \n",
+       "Rodrigo_Fill.1-Ariel_Guerreiro.1                 60.5003         0.830464  \n",
+       "...                                                  ...              ...  \n",
+       "Alex_Koji.1-Enzo_Bustos.1                        86.6502          1.00234  \n",
+       "Alex_Koji.1-Enzo_Bustos.2                        86.5941         0.923037  \n",
+       "Alex_Koji.1-Paulo_Sestini.1                      69.9926         0.911146  \n",
+       "Alex_Koji.1-Paulo_Sestini.2                       63.095           0.7912  \n",
+       "Alex_Koji.1-Alex_Koji.1                                0                0  \n",
+       "\n",
+       "[841 rows x 12 columns]"
+      ]
+     },
+     "execution_count": 7,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "metrics_df"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 8,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "Realizando o ensemble para cada par de indivíduos!: 100%|██████████| 841/841 [00:00<00:00, 2306.28it/s]\n"
+     ]
+    }
+   ],
+   "source": [
+    "verif_df = DeepFace.ensemble_dataframe(metrics_df)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 9,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/html": [
+       "<div>\n",
+       "<style scoped>\n",
+       "    .dataframe tbody tr th:only-of-type {\n",
+       "        vertical-align: middle;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe tbody tr th {\n",
+       "        vertical-align: top;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe thead th {\n",
+       "        text-align: right;\n",
+       "    }\n",
+       "</style>\n",
+       "<table border=\"1\" class=\"dataframe\">\n",
+       "  <thead>\n",
+       "    <tr style=\"text-align: right;\">\n",
+       "      <th></th>\n",
+       "      <th>Verified</th>\n",
+       "      <th>Reality</th>\n",
+       "      <th>Score</th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>Rodrigo_Fill.1-Rodrigo_Fill.1</th>\n",
+       "      <td>true</td>\n",
+       "      <td>true</td>\n",
+       "      <td>0.996973</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>Rodrigo_Fill.1-Rodrigo_Fill.2</th>\n",
+       "      <td>true</td>\n",
+       "      <td>true</td>\n",
+       "      <td>0.996922</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>Rodrigo_Fill.1-Luisa_Heise.1</th>\n",
+       "      <td>false</td>\n",
+       "      <td>false</td>\n",
+       "      <td>0.998384</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>Rodrigo_Fill.1-Luisa_Heise.2</th>\n",
+       "      <td>false</td>\n",
+       "      <td>false</td>\n",
+       "      <td>0.998695</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>Rodrigo_Fill.1-Ariel_Guerreiro.1</th>\n",
+       "      <td>false</td>\n",
+       "      <td>false</td>\n",
+       "      <td>0.998559</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>...</th>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "      <td>...</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>Alex_Koji.1-Enzo_Bustos.1</th>\n",
+       "      <td>false</td>\n",
+       "      <td>false</td>\n",
+       "      <td>0.998646</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>Alex_Koji.1-Enzo_Bustos.2</th>\n",
+       "      <td>false</td>\n",
+       "      <td>false</td>\n",
+       "      <td>0.998646</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>Alex_Koji.1-Paulo_Sestini.1</th>\n",
+       "      <td>false</td>\n",
+       "      <td>false</td>\n",
+       "      <td>0.998697</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>Alex_Koji.1-Paulo_Sestini.2</th>\n",
+       "      <td>false</td>\n",
+       "      <td>false</td>\n",
+       "      <td>0.998527</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>Alex_Koji.1-Alex_Koji.1</th>\n",
+       "      <td>true</td>\n",
+       "      <td>true</td>\n",
+       "      <td>0.996973</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "<p>841 rows × 3 columns</p>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "                                 Verified Reality     Score\n",
+       "Rodrigo_Fill.1-Rodrigo_Fill.1        true    true  0.996973\n",
+       "Rodrigo_Fill.1-Rodrigo_Fill.2        true    true  0.996922\n",
+       "Rodrigo_Fill.1-Luisa_Heise.1        false   false  0.998384\n",
+       "Rodrigo_Fill.1-Luisa_Heise.2        false   false  0.998695\n",
+       "Rodrigo_Fill.1-Ariel_Guerreiro.1    false   false  0.998559\n",
+       "...                                   ...     ...       ...\n",
+       "Alex_Koji.1-Enzo_Bustos.1           false   false  0.998646\n",
+       "Alex_Koji.1-Enzo_Bustos.2           false   false  0.998646\n",
+       "Alex_Koji.1-Paulo_Sestini.1         false   false  0.998697\n",
+       "Alex_Koji.1-Paulo_Sestini.2         false   false  0.998527\n",
+       "Alex_Koji.1-Alex_Koji.1              true    true  0.996973\n",
+       "\n",
+       "[841 rows x 3 columns]"
+      ]
+     },
+     "execution_count": 9,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "verif_df"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 10,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/html": [
+       "<div>\n",
+       "<style scoped>\n",
+       "    .dataframe tbody tr th:only-of-type {\n",
+       "        vertical-align: middle;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe tbody tr th {\n",
+       "        vertical-align: top;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe thead th {\n",
+       "        text-align: right;\n",
+       "    }\n",
+       "</style>\n",
+       "<table border=\"1\" class=\"dataframe\">\n",
+       "  <thead>\n",
+       "    <tr style=\"text-align: right;\">\n",
+       "      <th></th>\n",
+       "      <th>Accuracy</th>\n",
+       "      <th>Recall</th>\n",
+       "      <th>Precision</th>\n",
+       "      <th>F1</th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>Results:</th>\n",
+       "      <td>0.978597</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>0.814433</td>\n",
+       "      <td>0.897727</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "          Accuracy  Recall  Precision        F1\n",
+       "Results:  0.978597     1.0   0.814433  0.897727"
+      ]
+     },
+     "execution_count": 10,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "f1_df = DeepFace.f1_calculation(verif_df)\n",
+    "f1_df"
+   ]
   }
  ],
  "metadata": {
@@ -2348,7 +3402,7 @@
    "name": "python",
    "nbconvert_exporter": "python",
    "pygments_lexer": "ipython3",
-   "version": "3.6.8"
+   "version": "3.7.6"
   }
  },
  "nbformat": 4,
diff --git "a/Apresentacao_do_projeto/Apresenta\303\247\303\243o_Landmarks.ipynb" "b/Apresentacao_do_projeto/Apresenta\303\247\303\243o_Landmarks.ipynb"
index f47e578..139c462 100755
--- "a/Apresentacao_do_projeto/Apresenta\303\247\303\243o_Landmarks.ipynb"
+++ "b/Apresentacao_do_projeto/Apresenta\303\247\303\243o_Landmarks.ipynb"
@@ -1639,11 +1639,7 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "\n",
-    "\n",
-    "Sabendo então que Gui e Sib foram responsáveis por apresentar o grupo ao Big Boss, e sabendo que os resultados obtidos foram satisfatórios, em um intervado de tempo menor que a PUC. Assim é visível que ambos receberão bonus de desempenho, dependendo da ordem de redução dos custos é possível que eles possam pagar poke para o grupo por pelo menos um ano, todos os dias. Assim o próximo projeto é encontrar uma forma eficiente de cobrar nossos caros amigos :).\n",
-    "\n",
-    "aimport pickle\n",
+    "import pickle\n",
     "#save the model to disk\n",
     "filename = 'gradient_boosting_faces_clf.sav'\n",
     "pickle.dump(clf, open(filename, 'wb'))"