Demo.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "9abe16fb",
   "metadata": {},
   "source": [
    "Import Libraries"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "id": "b4324e8e",
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "from scipy.ndimage import gaussian_filter\n",
    "from PIL import Image\n",
    "from matplotlib.pyplot import specgram\n",
    "from matplotlib.pyplot import subplots\n",
    "from matplotlib.mlab import window_hanning\n",
    "import matplotlib.pyplot as plt\n",
    "import scipy\n",
    "import cv2\n",
    "%reload_ext autoreload\n",
    "%autoreload 2\n",
    "from fastai.imports import *\n",
    "from fastai.transforms import *\n",
    "from fastai.conv_learner import *\n",
    "from fastai.model import *\n",
    "from fastai.dataset import *\n",
    "from fastai.sgdr import *\n",
    "from fastai.plots import *\n",
    "from sklearn import metrics"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4e55b105",
   "metadata": {},
   "source": [
    "Import data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "id": "9ba4376a",
   "metadata": {},
   "outputs": [],
   "source": [
    "# data file\n",
    "dEDF = pd.read_csv('data.csv').values\n",
    "# time file\n",
    "tEDF = pd.read_csv('time.csv').values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "id": "228b4eca",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([100.])"
      ]
     },
     "execution_count": 36,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "max(tEDF)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "id": "f01e7dd5",
   "metadata": {},
   "outputs": [],
   "source": [
    "Fs = 1/(tEDF[3]-tEDF[2])\n",
    "Fs = Fs[0]\n",
    "winSize = int(np.round(Fs))\n",
    "winStep = int(np.round(Fs/2))\n",
    "movingwin    = [0.200,0.005] # Window size, Step size [s]\n",
    "iWin   = int(np.round(movingwin[0]*Fs/2)*2);\n",
    "iStep  = int(np.round(movingwin[1]*Fs/2)*2);\n",
    "noverlap = iWin-iStep\n",
    "i_stop = 0\n",
    "i = 0\n",
    "counter = 0"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "id": "c571b8a8",
   "metadata": {},
   "outputs": [],
   "source": [
    "def hannspecgramc(d1,movingwin,Fs,tEDF,flo,fhi):\n",
    "    iStart = 1;\n",
    "    iWin   = int(np.round(movingwin[0]*Fs/2)*2);\n",
    "    iStep  = int(np.round(movingwin[1]*Fs/2)*2);\n",
    "    T      = iWin/Fs;\n",
    "    df     = 1/T;\n",
    "    fNQ    = Fs/2;\n",
    "    freq   = range(0,int(iWin/2+1))*df;\n",
    "    counter=1;\n",
    "    S      = np.zeros([int(np.ceil(len(tEDF)/iStep)), int(iWin/2+1)]);\n",
    "    times  = np.zeros([int(np.ceil(len(tEDF)/iStep)), 1]);\n",
    "    while iStart+iWin < len(tEDF):\n",
    "        dnow    = d1[iStart:iStart+iWin-1];\n",
    "        dnow    = dnow - np.mean(dnow);\n",
    "        dnow    = np.hanning(len(dnow))*np.squeeze(dnow);\n",
    "        spectrum= np.real(scipy.fft.fft(dnow)*np.conjugate(scipy.fft.fft(dnow)));\n",
    "        S[counter,:]   = spectrum[:int(iWin/2+1)];\n",
    "        times[counter] = tEDF[int(iStart+iWin/2)];\n",
    "\n",
    "        counter=counter+1;\n",
    "        iStart = iStart + iStep;\n",
    "    S = S[:counter-1,:];\n",
    "    times = times[:counter-1];\n",
    "    S = S[:,(freq>=flo)&(freq<=fhi)]\n",
    "    freq = freq[(freq>=flo)&(freq<=fhi)]\n",
    "    return S,times,freq"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "id": "e8abb1f6",
   "metadata": {},
   "outputs": [],
   "source": [
    "start_times = []\n",
    "stop_times = []"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "id": "5ba370e4",
   "metadata": {},
   "outputs": [],
   "source": [
    "image_no_start_stop_dict = {}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "id": "9d0b5804",
   "metadata": {},
   "outputs": [],
   "source": [
    "while tEDF[i_stop] < 99: \n",
    "#only looking at the first K seconds, can change to any value\n",
    "\n",
    "    i_start = i\n",
    "    i_stop  = i_start + winSize - 1\n",
    "\n",
    "    dspec = dEDF[i_start:i_stop]\n",
    "    dspec = dspec - np.mean(dspec)\n",
    "    t     = tEDF[i_start:i_stop]\n",
    "    flo = 30\n",
    "    fhi = 250\n",
    "    movingwin    = [0.200,0.005] # Window size, Step size [s]\n",
    "    S0,S_times,S_freq = hannspecgramc(dspec,movingwin,Fs,t,flo,fhi)\n",
    "    #Smooth the spectra.\n",
    "    t_smooth = 11\n",
    "    dt_S     = S_times[2]-S_times[1]\n",
    "    S_smooth = gaussian_filter(S0,1)\n",
    "\n",
    "    A = np.log10(S_smooth)\n",
    "    A = np.flipud(A)\n",
    "    \n",
    "    #fig, ax = subplots(figsize=(6,6))\n",
    "    #ax.imshow(A, cmap='jet',\n",
    "    #       extent = [min(S_freq),max(S_freq),min(S_times)[0],max(S_times)[0]],\n",
    "    #         aspect = 'auto')\n",
    "    \n",
    "    name = 'demo_data/test/img_'+ str(counter)+'.jpg'\n",
    "    plt.imsave(name,A,cmap='jet')\n",
    "    \n",
    "    start_times.append(tEDF[i_start])\n",
    "    stop_times.append(tEDF[i_stop])\n",
    "    start_stop = [tEDF[i_start],tEDF[i_stop]]\n",
    "    \n",
    "    new_name = 'test/img_'+ str(counter)+'.jpg'\n",
    "    image_no_start_stop_dict[new_name] = start_stop\n",
    "    \n",
    "    i = i_start + winStep;\n",
    "    counter=counter+1;"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "id": "0f7084e7",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<All keys matched successfully>"
      ]
     },
     "execution_count": 42,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# where the data is\n",
    "PATH = \"demo_data\"\n",
    "# using resnet architecture\n",
    "arch = resnet34\n",
    "# size of square image in pixels\n",
    "sz = 44\n",
    "# transforms used on training data\n",
    "transforms_up_down = [RandomScale(sz,1.2),RandomRotate(1)]\n",
    "tfms = tfms_from_model(arch,sz,crop_type = CropType.NO,aug_tfms=transforms_up_down)\n",
    "# data: comes from PATH, used tfms on training data, bs of 8 for training data, test data located in test folder\n",
    "data = ImageClassifierData.from_paths(PATH,tfms=tfms,bs=8,test_name='test')\n",
    "# load in pretraine\n",
    "state = torch.load('saved_model.pkl',map_location=torch.device('cpu')) # remove map_location parameter if on GPU\n",
    "learn2 = ConvLearner.pretrained(arch,data,precompute=False)\n",
    "learn2.model.load_state_dict(state)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "id": "3c3cc4f6",
   "metadata": {},
   "outputs": [],
   "source": [
    "log_preds_test = learn2.predict(is_test=True)\n",
    "preds_test = np.argmax(log_preds_test,axis=1)\n",
    "probs_test = np.exp(log_preds_test[:,1])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "id": "9ba30c2c",
   "metadata": {},
   "outputs": [],
   "source": [
    "test_names = np.empty_like(data.test_ds.fnames)\n",
    "for i in range(len(data.test_ds.fnames)):\n",
    "    test_names[i] = data.test_ds.fnames[i]\n",
    "    #temp = data.test_ds.fnames[i]\n",
    "    #matchobj = re.search('.*im.*',temp)\n",
    "    #test_names[i] = matchobj.group()\n",
    "test_df = pd.DataFrame(data = test_names,columns = ['image_number'])\n",
    "test_df['prediction'] = preds_test\n",
    "test_df['probability'] = probs_test\n",
    "test_df['times'] = test_df['image_number'].map(image_no_start_stop_dict)\n",
    "#test_df['start time'] = start_times\n",
    "#test_df['stop time'] = stop_times"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "id": "21a63c7d",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'test/img_0.jpg': [array([0.00098]), array([1.00049])],\n",
       " 'test/img_1.jpg': [array([0.50074]), array([1.50025])],\n",
       " 'test/img_2.jpg': [array([1.00049]), array([2.])],\n",
       " 'test/img_3.jpg': [array([1.50025]), array([2.49975])],\n",
       " 'test/img_4.jpg': [array([2.]), array([2.99951])],\n",
       " 'test/img_5.jpg': [array([2.49975]), array([3.49926])],\n",
       " 'test/img_6.jpg': [array([2.99951]), array([3.99902])],\n",
       " 'test/img_7.jpg': [array([3.49926]), array([4.49877])],\n",
       " 'test/img_8.jpg': [array([3.99902]), array([4.99853])],\n",
       " 'test/img_9.jpg': [array([4.49877]), array([5.49828])],\n",
       " 'test/img_10.jpg': [array([4.99853]), array([5.99803])],\n",
       " 'test/img_11.jpg': [array([5.49828]), array([6.49779])],\n",
       " 'test/img_12.jpg': [array([5.99803]), array([6.99754])],\n",
       " 'test/img_13.jpg': [array([6.49779]), array([7.4973])],\n",
       " 'test/img_14.jpg': [array([6.99754]), array([7.99705])],\n",
       " 'test/img_15.jpg': [array([7.4973]), array([8.49681])],\n",
       " 'test/img_16.jpg': [array([7.99705]), array([8.99656])],\n",
       " 'test/img_17.jpg': [array([8.49681]), array([9.49631])],\n",
       " 'test/img_18.jpg': [array([8.99656]), array([9.99607])],\n",
       " 'test/img_19.jpg': [array([9.49631]), array([10.49582])],\n",
       " 'test/img_20.jpg': [array([9.99607]), array([10.99558])],\n",
       " 'test/img_21.jpg': [array([10.49582]), array([11.49533])],\n",
       " 'test/img_22.jpg': [array([10.99558]), array([11.99509])],\n",
       " 'test/img_23.jpg': [array([11.49533]), array([12.49484])],\n",
       " 'test/img_24.jpg': [array([11.99509]), array([12.99459])],\n",
       " 'test/img_25.jpg': [array([12.49484]), array([13.49435])],\n",
       " 'test/img_26.jpg': [array([12.99459]), array([13.9941])],\n",
       " 'test/img_27.jpg': [array([13.49435]), array([14.49386])],\n",
       " 'test/img_28.jpg': [array([13.9941]), array([14.99361])],\n",
       " 'test/img_29.jpg': [array([14.49386]), array([15.49337])],\n",
       " 'test/img_30.jpg': [array([14.99361]), array([15.99312])],\n",
       " 'test/img_31.jpg': [array([15.49337]), array([16.49287])],\n",
       " 'test/img_32.jpg': [array([15.99312]), array([16.99263])],\n",
       " 'test/img_33.jpg': [array([16.49287]), array([17.49238])],\n",
       " 'test/img_34.jpg': [array([16.99263]), array([17.99214])],\n",
       " 'test/img_35.jpg': [array([17.49238]), array([18.49189])],\n",
       " 'test/img_36.jpg': [array([17.99214]), array([18.99165])],\n",
       " 'test/img_37.jpg': [array([18.49189]), array([19.4914])],\n",
       " 'test/img_38.jpg': [array([18.99165]), array([19.99115])],\n",
       " 'test/img_39.jpg': [array([19.4914]), array([20.49091])],\n",
       " 'test/img_40.jpg': [array([19.99115]), array([20.99066])],\n",
       " 'test/img_41.jpg': [array([20.49091]), array([21.49042])],\n",
       " 'test/img_42.jpg': [array([20.99066]), array([21.99017])],\n",
       " 'test/img_43.jpg': [array([21.49042]), array([22.48993])],\n",
       " 'test/img_44.jpg': [array([21.99017]), array([22.98968])],\n",
       " 'test/img_45.jpg': [array([22.48993]), array([23.48943])],\n",
       " 'test/img_46.jpg': [array([22.98968]), array([23.98919])],\n",
       " 'test/img_47.jpg': [array([23.48943]), array([24.48894])],\n",
       " 'test/img_48.jpg': [array([23.98919]), array([24.9887])],\n",
       " 'test/img_49.jpg': [array([24.48894]), array([25.48845])],\n",
       " 'test/img_50.jpg': [array([24.9887]), array([25.98821])],\n",
       " 'test/img_51.jpg': [array([25.48845]), array([26.48796])],\n",
       " 'test/img_52.jpg': [array([25.98821]), array([26.98771])],\n",
       " 'test/img_53.jpg': [array([26.48796]), array([27.48747])],\n",
       " 'test/img_54.jpg': [array([26.98771]), array([27.98722])],\n",
       " 'test/img_55.jpg': [array([27.48747]), array([28.48698])],\n",
       " 'test/img_56.jpg': [array([27.98722]), array([28.98673])],\n",
       " 'test/img_57.jpg': [array([28.48698]), array([29.48649])],\n",
       " 'test/img_58.jpg': [array([28.98673]), array([29.98624])],\n",
       " 'test/img_59.jpg': [array([29.48649]), array([30.486])],\n",
       " 'test/img_60.jpg': [array([29.98624]), array([30.98575])],\n",
       " 'test/img_61.jpg': [array([30.486]), array([31.4855])],\n",
       " 'test/img_62.jpg': [array([30.98575]), array([31.98526])],\n",
       " 'test/img_63.jpg': [array([31.4855]), array([32.48501])],\n",
       " 'test/img_64.jpg': [array([31.98526]), array([32.98477])],\n",
       " 'test/img_65.jpg': [array([32.48501]), array([33.48452])],\n",
       " 'test/img_66.jpg': [array([32.98477]), array([33.98428])],\n",
       " 'test/img_67.jpg': [array([33.48452]), array([34.48403])],\n",
       " 'test/img_68.jpg': [array([33.98428]), array([34.98378])],\n",
       " 'test/img_69.jpg': [array([34.48403]), array([35.48354])],\n",
       " 'test/img_70.jpg': [array([34.98378]), array([35.98329])],\n",
       " 'test/img_71.jpg': [array([35.48354]), array([36.48305])],\n",
       " 'test/img_72.jpg': [array([35.98329]), array([36.9828])],\n",
       " 'test/img_73.jpg': [array([36.48305]), array([37.48256])],\n",
       " 'test/img_74.jpg': [array([36.9828]), array([37.98231])],\n",
       " 'test/img_75.jpg': [array([37.48256]), array([38.48206])],\n",
       " 'test/img_76.jpg': [array([37.98231]), array([38.98182])],\n",
       " 'test/img_77.jpg': [array([38.48206]), array([39.48157])],\n",
       " 'test/img_78.jpg': [array([38.98182]), array([39.98133])],\n",
       " 'test/img_79.jpg': [array([39.48157]), array([40.48108])],\n",
       " 'test/img_80.jpg': [array([39.98133]), array([40.98084])],\n",
       " 'test/img_81.jpg': [array([40.48108]), array([41.48059])],\n",
       " 'test/img_82.jpg': [array([40.98084]), array([41.98034])],\n",
       " 'test/img_83.jpg': [array([41.48059]), array([42.4801])],\n",
       " 'test/img_84.jpg': [array([41.98034]), array([42.97985])],\n",
       " 'test/img_85.jpg': [array([42.4801]), array([43.47961])],\n",
       " 'test/img_86.jpg': [array([42.97985]), array([43.97936])],\n",
       " 'test/img_87.jpg': [array([43.47961]), array([44.47912])],\n",
       " 'test/img_88.jpg': [array([43.97936]), array([44.97887])],\n",
       " 'test/img_89.jpg': [array([44.47912]), array([45.47862])],\n",
       " 'test/img_90.jpg': [array([44.97887]), array([45.97838])],\n",
       " 'test/img_91.jpg': [array([45.47862]), array([46.47813])],\n",
       " 'test/img_92.jpg': [array([45.97838]), array([46.97789])],\n",
       " 'test/img_93.jpg': [array([46.47813]), array([47.47764])],\n",
       " 'test/img_94.jpg': [array([46.97789]), array([47.9774])],\n",
       " 'test/img_95.jpg': [array([47.47764]), array([48.47715])],\n",
       " 'test/img_96.jpg': [array([47.9774]), array([48.9769])],\n",
       " 'test/img_97.jpg': [array([48.47715]), array([49.47666])],\n",
       " 'test/img_98.jpg': [array([48.9769]), array([49.97641])],\n",
       " 'test/img_99.jpg': [array([49.47666]), array([50.47617])],\n",
       " 'test/img_100.jpg': [array([49.97641]), array([50.97592])],\n",
       " 'test/img_101.jpg': [array([50.47617]), array([51.47568])],\n",
       " 'test/img_102.jpg': [array([50.97592]), array([51.97543])],\n",
       " 'test/img_103.jpg': [array([51.47568]), array([52.47518])],\n",
       " 'test/img_104.jpg': [array([51.97543]), array([52.97494])],\n",
       " 'test/img_105.jpg': [array([52.47518]), array([53.47469])],\n",
       " 'test/img_106.jpg': [array([52.97494]), array([53.97445])],\n",
       " 'test/img_107.jpg': [array([53.47469]), array([54.4742])],\n",
       " 'test/img_108.jpg': [array([53.97445]), array([54.97396])],\n",
       " 'test/img_109.jpg': [array([54.4742]), array([55.47371])],\n",
       " 'test/img_110.jpg': [array([54.97396]), array([55.97346])],\n",
       " 'test/img_111.jpg': [array([55.47371]), array([56.47322])],\n",
       " 'test/img_112.jpg': [array([55.97346]), array([56.97297])],\n",
       " 'test/img_113.jpg': [array([56.47322]), array([57.47273])],\n",
       " 'test/img_114.jpg': [array([56.97297]), array([57.97248])],\n",
       " 'test/img_115.jpg': [array([57.47273]), array([58.47224])],\n",
       " 'test/img_116.jpg': [array([57.97248]), array([58.97199])],\n",
       " 'test/img_117.jpg': [array([58.47224]), array([59.47174])],\n",
       " 'test/img_118.jpg': [array([58.97199]), array([59.9715])],\n",
       " 'test/img_119.jpg': [array([59.47174]), array([60.47125])],\n",
       " 'test/img_120.jpg': [array([59.9715]), array([60.97101])],\n",
       " 'test/img_121.jpg': [array([60.47125]), array([61.47076])],\n",
       " 'test/img_122.jpg': [array([60.97101]), array([61.97052])],\n",
       " 'test/img_123.jpg': [array([61.47076]), array([62.47027])],\n",
       " 'test/img_124.jpg': [array([61.97052]), array([62.97002])],\n",
       " 'test/img_125.jpg': [array([62.47027]), array([63.46978])],\n",
       " 'test/img_126.jpg': [array([62.97002]), array([63.96953])],\n",
       " 'test/img_127.jpg': [array([63.46978]), array([64.46929])],\n",
       " 'test/img_128.jpg': [array([63.96953]), array([64.96904])],\n",
       " 'test/img_129.jpg': [array([64.46929]), array([65.4688])],\n",
       " 'test/img_130.jpg': [array([64.96904]), array([65.96855])],\n",
       " 'test/img_131.jpg': [array([65.4688]), array([66.4683])],\n",
       " 'test/img_132.jpg': [array([65.96855]), array([66.96806])],\n",
       " 'test/img_133.jpg': [array([66.4683]), array([67.46781])],\n",
       " 'test/img_134.jpg': [array([66.96806]), array([67.96757])],\n",
       " 'test/img_135.jpg': [array([67.46781]), array([68.46732])],\n",
       " 'test/img_136.jpg': [array([67.96757]), array([68.96708])],\n",
       " 'test/img_137.jpg': [array([68.46732]), array([69.46683])],\n",
       " 'test/img_138.jpg': [array([68.96708]), array([69.96658])],\n",
       " 'test/img_139.jpg': [array([69.46683]), array([70.46634])],\n",
       " 'test/img_140.jpg': [array([69.96658]), array([70.96609])],\n",
       " 'test/img_141.jpg': [array([70.46634]), array([71.46585])],\n",
       " 'test/img_142.jpg': [array([70.96609]), array([71.9656])],\n",
       " 'test/img_143.jpg': [array([71.46585]), array([72.46536])],\n",
       " 'test/img_144.jpg': [array([71.9656]), array([72.96511])],\n",
       " 'test/img_145.jpg': [array([72.46536]), array([73.46486])],\n",
       " 'test/img_146.jpg': [array([72.96511]), array([73.96462])],\n",
       " 'test/img_147.jpg': [array([73.46486]), array([74.46437])],\n",
       " 'test/img_148.jpg': [array([73.96462]), array([74.96413])],\n",
       " 'test/img_149.jpg': [array([74.46437]), array([75.46388])],\n",
       " 'test/img_150.jpg': [array([74.96413]), array([75.96364])],\n",
       " 'test/img_151.jpg': [array([75.46388]), array([76.46339])],\n",
       " 'test/img_152.jpg': [array([75.96364]), array([76.96314])],\n",
       " 'test/img_153.jpg': [array([76.46339]), array([77.4629])],\n",
       " 'test/img_154.jpg': [array([76.96314]), array([77.96265])],\n",
       " 'test/img_155.jpg': [array([77.4629]), array([78.46241])],\n",
       " 'test/img_156.jpg': [array([77.96265]), array([78.96216])],\n",
       " 'test/img_157.jpg': [array([78.46241]), array([79.46192])],\n",
       " 'test/img_158.jpg': [array([78.96216]), array([79.96167])],\n",
       " 'test/img_159.jpg': [array([79.46192]), array([80.46143])],\n",
       " 'test/img_160.jpg': [array([79.96167]), array([80.96118])],\n",
       " 'test/img_161.jpg': [array([80.46143]), array([81.46093])],\n",
       " 'test/img_162.jpg': [array([80.96118]), array([81.96069])],\n",
       " 'test/img_163.jpg': [array([81.46093]), array([82.46044])],\n",
       " 'test/img_164.jpg': [array([81.96069]), array([82.9602])],\n",
       " 'test/img_165.jpg': [array([82.46044]), array([83.45995])],\n",
       " 'test/img_166.jpg': [array([82.9602]), array([83.95971])],\n",
       " 'test/img_167.jpg': [array([83.45995]), array([84.45946])],\n",
       " 'test/img_168.jpg': [array([83.95971]), array([84.95921])],\n",
       " 'test/img_169.jpg': [array([84.45946]), array([85.45897])],\n",
       " 'test/img_170.jpg': [array([84.95921]), array([85.95872])],\n",
       " 'test/img_171.jpg': [array([85.45897]), array([86.45848])],\n",
       " 'test/img_172.jpg': [array([85.95872]), array([86.95823])],\n",
       " 'test/img_173.jpg': [array([86.45848]), array([87.45799])],\n",
       " 'test/img_174.jpg': [array([86.95823]), array([87.95774])],\n",
       " 'test/img_175.jpg': [array([87.45799]), array([88.45749])],\n",
       " 'test/img_176.jpg': [array([87.95774]), array([88.95725])],\n",
       " 'test/img_177.jpg': [array([88.45749]), array([89.457])],\n",
       " 'test/img_178.jpg': [array([88.95725]), array([89.95676])],\n",
       " 'test/img_179.jpg': [array([89.457]), array([90.45651])],\n",
       " 'test/img_180.jpg': [array([89.95676]), array([90.95627])],\n",
       " 'test/img_181.jpg': [array([90.45651]), array([91.45602])],\n",
       " 'test/img_182.jpg': [array([90.95627]), array([91.95577])],\n",
       " 'test/img_183.jpg': [array([91.45602]), array([92.45553])],\n",
       " 'test/img_184.jpg': [array([91.95577]), array([92.95528])],\n",
       " 'test/img_185.jpg': [array([92.45553]), array([93.45504])],\n",
       " 'test/img_186.jpg': [array([92.95528]), array([93.95479])],\n",
       " 'test/img_187.jpg': [array([93.45504]), array([94.45455])],\n",
       " 'test/img_188.jpg': [array([93.95479]), array([94.9543])],\n",
       " 'test/img_189.jpg': [array([94.45455]), array([95.45405])],\n",
       " 'test/img_190.jpg': [array([94.9543]), array([95.95381])],\n",
       " 'test/img_191.jpg': [array([95.45405]), array([96.45356])],\n",
       " 'test/img_192.jpg': [array([95.95381]), array([96.95332])],\n",
       " 'test/img_193.jpg': [array([96.45356]), array([97.45307])],\n",
       " 'test/img_194.jpg': [array([96.95332]), array([97.95283])],\n",
       " 'test/img_195.jpg': [array([97.45307]), array([98.45258])],\n",
       " 'test/img_196.jpg': [array([97.95283]), array([98.95233])],\n",
       " 'test/img_197.jpg': [array([98.45258]), array([99.45209])]}"
      ]
     },
     "execution_count": 45,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "image_no_start_stop_dict"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "id": "7ebdbc53",
   "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>image_number</th>\n",
       "      <th>prediction</th>\n",
       "      <th>probability</th>\n",
       "      <th>times</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>136</th>\n",
       "      <td>test/img_41.jpg</td>\n",
       "      <td>1</td>\n",
       "      <td>0.625198</td>\n",
       "      <td>[[20.4909090909091], [21.4904176904177]]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>93</th>\n",
       "      <td>test/img_196.jpg</td>\n",
       "      <td>1</td>\n",
       "      <td>0.675399</td>\n",
       "      <td>[[97.9528255528256], [98.9523341523341]]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>84</th>\n",
       "      <td>test/img_77.jpg</td>\n",
       "      <td>1</td>\n",
       "      <td>0.753126</td>\n",
       "      <td>[[38.4820638820639], [39.4815724815725]]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>69</th>\n",
       "      <td>test/img_173.jpg</td>\n",
       "      <td>1</td>\n",
       "      <td>0.774374</td>\n",
       "      <td>[[86.4584766584767], [87.4579852579853]]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>186</th>\n",
       "      <td>test/img_131.jpg</td>\n",
       "      <td>1</td>\n",
       "      <td>0.841441</td>\n",
       "      <td>[[65.4687960687961], [66.4683046683047]]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>129</th>\n",
       "      <td>test/img_184.jpg</td>\n",
       "      <td>1</td>\n",
       "      <td>0.899287</td>\n",
       "      <td>[[91.9557739557739], [92.9552825552826]]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>139</th>\n",
       "      <td>test/img_82.jpg</td>\n",
       "      <td>1</td>\n",
       "      <td>0.918211</td>\n",
       "      <td>[[40.9808353808354], [41.980343980344]]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>180</th>\n",
       "      <td>test/img_23.jpg</td>\n",
       "      <td>1</td>\n",
       "      <td>0.920939</td>\n",
       "      <td>[[11.4953316953317], [12.4948402948403]]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>37</th>\n",
       "      <td>test/img_113.jpg</td>\n",
       "      <td>1</td>\n",
       "      <td>0.928052</td>\n",
       "      <td>[[56.4732186732187], [57.4727272727273]]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>48</th>\n",
       "      <td>test/img_149.jpg</td>\n",
       "      <td>1</td>\n",
       "      <td>0.934602</td>\n",
       "      <td>[[74.4643734643735], [75.4638820638821]]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>109</th>\n",
       "      <td>test/img_142.jpg</td>\n",
       "      <td>1</td>\n",
       "      <td>0.940355</td>\n",
       "      <td>[[70.9660933660934], [71.965601965602]]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>132</th>\n",
       "      <td>test/img_185.jpg</td>\n",
       "      <td>1</td>\n",
       "      <td>0.947337</td>\n",
       "      <td>[[92.4555282555283], [93.4550368550369]]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75</th>\n",
       "      <td>test/img_76.jpg</td>\n",
       "      <td>1</td>\n",
       "      <td>0.959809</td>\n",
       "      <td>[[37.9823095823096], [38.9818181818182]]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>122</th>\n",
       "      <td>test/img_83.jpg</td>\n",
       "      <td>1</td>\n",
       "      <td>0.964388</td>\n",
       "      <td>[[41.4805896805897], [42.4800982800983]]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>32</th>\n",
       "      <td>test/img_112.jpg</td>\n",
       "      <td>1</td>\n",
       "      <td>0.965894</td>\n",
       "      <td>[[55.9734643734644], [56.972972972973]]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>125</th>\n",
       "      <td>test/img_40.jpg</td>\n",
       "      <td>1</td>\n",
       "      <td>0.970302</td>\n",
       "      <td>[[19.9911547911548], [20.9906633906634]]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>98</th>\n",
       "      <td>test/img_154.jpg</td>\n",
       "      <td>1</td>\n",
       "      <td>0.971112</td>\n",
       "      <td>[[76.963144963145], [77.9626535626536]]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>116</th>\n",
       "      <td>test/img_143.jpg</td>\n",
       "      <td>1</td>\n",
       "      <td>0.984869</td>\n",
       "      <td>[[71.4658476658477], [72.4653562653563]]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>89</th>\n",
       "      <td>test/img_46.jpg</td>\n",
       "      <td>1</td>\n",
       "      <td>0.987878</td>\n",
       "      <td>[[22.9896805896806], [23.9891891891892]]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>61</th>\n",
       "      <td>test/img_59.jpg</td>\n",
       "      <td>1</td>\n",
       "      <td>0.988274</td>\n",
       "      <td>[[29.4864864864865], [30.4859950859951]]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>130</th>\n",
       "      <td>test/img_190.jpg</td>\n",
       "      <td>1</td>\n",
       "      <td>0.988684</td>\n",
       "      <td>[[94.9542997542998], [95.9538083538084]]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>187</th>\n",
       "      <td>test/img_125.jpg</td>\n",
       "      <td>1</td>\n",
       "      <td>0.989037</td>\n",
       "      <td>[[62.4702702702703], [63.4697788697789]]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>197</th>\n",
       "      <td>test/img_35.jpg</td>\n",
       "      <td>1</td>\n",
       "      <td>0.989394</td>\n",
       "      <td>[[17.4923832923833], [18.4918918918919]]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>131</th>\n",
       "      <td>test/img_191.jpg</td>\n",
       "      <td>1</td>\n",
       "      <td>0.989962</td>\n",
       "      <td>[[95.4540540540541], [96.4535626535627]]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>test/img_100.jpg</td>\n",
       "      <td>1</td>\n",
       "      <td>0.990982</td>\n",
       "      <td>[[49.9764127764128], [50.9759213759214]]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>45</th>\n",
       "      <td>test/img_160.jpg</td>\n",
       "      <td>1</td>\n",
       "      <td>0.991070</td>\n",
       "      <td>[[79.9616707616708], [80.9611793611794]]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>46</th>\n",
       "      <td>test/img_161.jpg</td>\n",
       "      <td>1</td>\n",
       "      <td>0.991791</td>\n",
       "      <td>[[80.4614250614251], [81.4609336609337]]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>96</th>\n",
       "      <td>test/img_197.jpg</td>\n",
       "      <td>1</td>\n",
       "      <td>0.992327</td>\n",
       "      <td>[[98.4525798525799], [99.4520884520884]]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>161</th>\n",
       "      <td>test/img_136.jpg</td>\n",
       "      <td>1</td>\n",
       "      <td>0.992416</td>\n",
       "      <td>[[67.9675675675676], [68.9670761670762]]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>43</th>\n",
       "      <td>test/img_148.jpg</td>\n",
       "      <td>1</td>\n",
       "      <td>0.992856</td>\n",
       "      <td>[[73.9646191646192], [74.9641277641278]]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>100</th>\n",
       "      <td>test/img_47.jpg</td>\n",
       "      <td>1</td>\n",
       "      <td>0.993352</td>\n",
       "      <td>[[23.4894348894349], [24.4889434889435]]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>36</th>\n",
       "      <td>test/img_107.jpg</td>\n",
       "      <td>1</td>\n",
       "      <td>0.993405</td>\n",
       "      <td>[[53.4746928746929], [54.4742014742015]]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>140</th>\n",
       "      <td>test/img_94.jpg</td>\n",
       "      <td>1</td>\n",
       "      <td>0.993716</td>\n",
       "      <td>[[46.977886977887], [47.9773955773956]]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>189</th>\n",
       "      <td>test/img_22.jpg</td>\n",
       "      <td>1</td>\n",
       "      <td>0.994381</td>\n",
       "      <td>[[10.9955773955774], [11.995085995086]]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>70</th>\n",
       "      <td>test/img_167.jpg</td>\n",
       "      <td>1</td>\n",
       "      <td>0.995368</td>\n",
       "      <td>[[83.4599508599509], [84.4594594594595]]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>182</th>\n",
       "      <td>test/img_124.jpg</td>\n",
       "      <td>1</td>\n",
       "      <td>0.995536</td>\n",
       "      <td>[[61.970515970516], [62.9700245700246]]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>33</th>\n",
       "      <td>test/img_106.jpg</td>\n",
       "      <td>1</td>\n",
       "      <td>0.996174</td>\n",
       "      <td>[[52.9749385749386], [53.9744471744472]]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20</th>\n",
       "      <td>test/img_16.jpg</td>\n",
       "      <td>1</td>\n",
       "      <td>0.997046</td>\n",
       "      <td>[[7.9970515970516], [8.9965601965602]]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>190</th>\n",
       "      <td>test/img_34.jpg</td>\n",
       "      <td>1</td>\n",
       "      <td>0.997087</td>\n",
       "      <td>[[16.992628992629], [17.9921375921376]]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>67</th>\n",
       "      <td>test/img_166.jpg</td>\n",
       "      <td>1</td>\n",
       "      <td>0.997455</td>\n",
       "      <td>[[82.9601965601966], [83.9597051597052]]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>53</th>\n",
       "      <td>test/img_70.jpg</td>\n",
       "      <td>1</td>\n",
       "      <td>0.997456</td>\n",
       "      <td>[[34.9837837837838], [35.9832923832924]]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>91</th>\n",
       "      <td>test/img_155.jpg</td>\n",
       "      <td>1</td>\n",
       "      <td>0.997477</td>\n",
       "      <td>[[77.4628992628993], [78.4624078624079]]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>54</th>\n",
       "      <td>test/img_58.jpg</td>\n",
       "      <td>1</td>\n",
       "      <td>0.997539</td>\n",
       "      <td>[[28.9867321867322], [29.9862407862408]]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>183</th>\n",
       "      <td>test/img_130.jpg</td>\n",
       "      <td>1</td>\n",
       "      <td>0.998203</td>\n",
       "      <td>[[64.9690417690418], [65.9685503685504]]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>39</th>\n",
       "      <td>test/img_28.jpg</td>\n",
       "      <td>1</td>\n",
       "      <td>0.998310</td>\n",
       "      <td>[[13.9941031941032], [14.9936117936118]]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>157</th>\n",
       "      <td>test/img_95.jpg</td>\n",
       "      <td>1</td>\n",
       "      <td>0.998321</td>\n",
       "      <td>[[47.4776412776413], [48.4771498771499]]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>88</th>\n",
       "      <td>test/img_52.jpg</td>\n",
       "      <td>1</td>\n",
       "      <td>0.999293</td>\n",
       "      <td>[[25.9882063882064], [26.987714987715]]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>184</th>\n",
       "      <td>test/img_118.jpg</td>\n",
       "      <td>1</td>\n",
       "      <td>0.999319</td>\n",
       "      <td>[[58.9719901719902], [59.9714987714988]]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>185</th>\n",
       "      <td>test/img_119.jpg</td>\n",
       "      <td>1</td>\n",
       "      <td>0.999669</td>\n",
       "      <td>[[59.4717444717445], [60.4712530712531]]</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "         image_number  prediction  probability  \\\n",
       "136   test/img_41.jpg           1     0.625198   \n",
       "93   test/img_196.jpg           1     0.675399   \n",
       "84    test/img_77.jpg           1     0.753126   \n",
       "69   test/img_173.jpg           1     0.774374   \n",
       "186  test/img_131.jpg           1     0.841441   \n",
       "129  test/img_184.jpg           1     0.899287   \n",
       "139   test/img_82.jpg           1     0.918211   \n",
       "180   test/img_23.jpg           1     0.920939   \n",
       "37   test/img_113.jpg           1     0.928052   \n",
       "48   test/img_149.jpg           1     0.934602   \n",
       "109  test/img_142.jpg           1     0.940355   \n",
       "132  test/img_185.jpg           1     0.947337   \n",
       "75    test/img_76.jpg           1     0.959809   \n",
       "122   test/img_83.jpg           1     0.964388   \n",
       "32   test/img_112.jpg           1     0.965894   \n",
       "125   test/img_40.jpg           1     0.970302   \n",
       "98   test/img_154.jpg           1     0.971112   \n",
       "116  test/img_143.jpg           1     0.984869   \n",
       "89    test/img_46.jpg           1     0.987878   \n",
       "61    test/img_59.jpg           1     0.988274   \n",
       "130  test/img_190.jpg           1     0.988684   \n",
       "187  test/img_125.jpg           1     0.989037   \n",
       "197   test/img_35.jpg           1     0.989394   \n",
       "131  test/img_191.jpg           1     0.989962   \n",
       "12   test/img_100.jpg           1     0.990982   \n",
       "45   test/img_160.jpg           1     0.991070   \n",
       "46   test/img_161.jpg           1     0.991791   \n",
       "96   test/img_197.jpg           1     0.992327   \n",
       "161  test/img_136.jpg           1     0.992416   \n",
       "43   test/img_148.jpg           1     0.992856   \n",
       "100   test/img_47.jpg           1     0.993352   \n",
       "36   test/img_107.jpg           1     0.993405   \n",
       "140   test/img_94.jpg           1     0.993716   \n",
       "189   test/img_22.jpg           1     0.994381   \n",
       "70   test/img_167.jpg           1     0.995368   \n",
       "182  test/img_124.jpg           1     0.995536   \n",
       "33   test/img_106.jpg           1     0.996174   \n",
       "20    test/img_16.jpg           1     0.997046   \n",
       "190   test/img_34.jpg           1     0.997087   \n",
       "67   test/img_166.jpg           1     0.997455   \n",
       "53    test/img_70.jpg           1     0.997456   \n",
       "91   test/img_155.jpg           1     0.997477   \n",
       "54    test/img_58.jpg           1     0.997539   \n",
       "183  test/img_130.jpg           1     0.998203   \n",
       "39    test/img_28.jpg           1     0.998310   \n",
       "157   test/img_95.jpg           1     0.998321   \n",
       "88    test/img_52.jpg           1     0.999293   \n",
       "184  test/img_118.jpg           1     0.999319   \n",
       "185  test/img_119.jpg           1     0.999669   \n",
       "\n",
       "                                        times  \n",
       "136  [[20.4909090909091], [21.4904176904177]]  \n",
       "93   [[97.9528255528256], [98.9523341523341]]  \n",
       "84   [[38.4820638820639], [39.4815724815725]]  \n",
       "69   [[86.4584766584767], [87.4579852579853]]  \n",
       "186  [[65.4687960687961], [66.4683046683047]]  \n",
       "129  [[91.9557739557739], [92.9552825552826]]  \n",
       "139   [[40.9808353808354], [41.980343980344]]  \n",
       "180  [[11.4953316953317], [12.4948402948403]]  \n",
       "37   [[56.4732186732187], [57.4727272727273]]  \n",
       "48   [[74.4643734643735], [75.4638820638821]]  \n",
       "109   [[70.9660933660934], [71.965601965602]]  \n",
       "132  [[92.4555282555283], [93.4550368550369]]  \n",
       "75   [[37.9823095823096], [38.9818181818182]]  \n",
       "122  [[41.4805896805897], [42.4800982800983]]  \n",
       "32    [[55.9734643734644], [56.972972972973]]  \n",
       "125  [[19.9911547911548], [20.9906633906634]]  \n",
       "98    [[76.963144963145], [77.9626535626536]]  \n",
       "116  [[71.4658476658477], [72.4653562653563]]  \n",
       "89   [[22.9896805896806], [23.9891891891892]]  \n",
       "61   [[29.4864864864865], [30.4859950859951]]  \n",
       "130  [[94.9542997542998], [95.9538083538084]]  \n",
       "187  [[62.4702702702703], [63.4697788697789]]  \n",
       "197  [[17.4923832923833], [18.4918918918919]]  \n",
       "131  [[95.4540540540541], [96.4535626535627]]  \n",
       "12   [[49.9764127764128], [50.9759213759214]]  \n",
       "45   [[79.9616707616708], [80.9611793611794]]  \n",
       "46   [[80.4614250614251], [81.4609336609337]]  \n",
       "96   [[98.4525798525799], [99.4520884520884]]  \n",
       "161  [[67.9675675675676], [68.9670761670762]]  \n",
       "43   [[73.9646191646192], [74.9641277641278]]  \n",
       "100  [[23.4894348894349], [24.4889434889435]]  \n",
       "36   [[53.4746928746929], [54.4742014742015]]  \n",
       "140   [[46.977886977887], [47.9773955773956]]  \n",
       "189   [[10.9955773955774], [11.995085995086]]  \n",
       "70   [[83.4599508599509], [84.4594594594595]]  \n",
       "182   [[61.970515970516], [62.9700245700246]]  \n",
       "33   [[52.9749385749386], [53.9744471744472]]  \n",
       "20     [[7.9970515970516], [8.9965601965602]]  \n",
       "190   [[16.992628992629], [17.9921375921376]]  \n",
       "67   [[82.9601965601966], [83.9597051597052]]  \n",
       "53   [[34.9837837837838], [35.9832923832924]]  \n",
       "91   [[77.4628992628993], [78.4624078624079]]  \n",
       "54   [[28.9867321867322], [29.9862407862408]]  \n",
       "183  [[64.9690417690418], [65.9685503685504]]  \n",
       "39   [[13.9941031941032], [14.9936117936118]]  \n",
       "157  [[47.4776412776413], [48.4771498771499]]  \n",
       "88    [[25.9882063882064], [26.987714987715]]  \n",
       "184  [[58.9719901719902], [59.9714987714988]]  \n",
       "185  [[59.4717444717445], [60.4712530712531]]  "
      ]
     },
     "execution_count": 46,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test_df[test_df['prediction']==1].sort_values('probability')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "0692fa4b",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "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": "3.7.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}