└── Class Activation Maps without MaxPooling.ipynb


/Class Activation Maps without MaxPooling.ipynb:
--------------------------------------------------------------------------------
  1 | {
  2 |  "cells": [
  3 |   {
  4 |    "cell_type": "code",
  5 |    "execution_count": 1,
  6 |    "metadata": {},
  7 |    "outputs": [
  8 |     {
  9 |      "name": "stderr",
 10 |      "output_type": "stream",
 11 |      "text": [
 12 |       "Using TensorFlow backend.\n"
 13 |      ]
 14 |     }
 15 |    ],
 16 |    "source": [
 17 |     "import keras\n",
 18 |     "from keras.datasets import mnist\n",
 19 |     "import numpy as np\n",
 20 |     "import matplotlib.pyplot as plt\n",
 21 |     "from keras.models import Sequential,Model\n",
 22 |     "from keras.layers import Dense,Conv2D,Flatten,MaxPooling2D,GlobalAveragePooling2D\n",
 23 |     "from keras.utils import plot_model"
 24 |    ]
 25 |   },
 26 |   {
 27 |    "cell_type": "code",
 28 |    "execution_count": 2,
 29 |    "metadata": {},
 30 |    "outputs": [
 31 |     {
 32 |      "name": "stdout",
 33 |      "output_type": "stream",
 34 |      "text": [
 35 |       "(60000, 28, 28)\n"
 36 |      ]
 37 |     }
 38 |    ],
 39 |    "source": [
 40 |     "(X_train,Y_train),(X_test,Y_test)  = mnist.load_data()\n",
 41 |     "print(X_train.shape)"
 42 |    ]
 43 |   },
 44 |   {
 45 |    "cell_type": "code",
 46 |    "execution_count": 3,
 47 |    "metadata": {},
 48 |    "outputs": [
 49 |     {
 50 |      "name": "stdout",
 51 |      "output_type": "stream",
 52 |      "text": [
 53 |       "(60000, 28, 28, 1)\n",
 54 |       "(10000, 28, 28, 1)\n"
 55 |      ]
 56 |     },
 57 |     {
 58 |      "data": {
 59 |       "text/plain": [
 60 |        "(28, 28, 1)"
 61 |       ]
 62 |      },
 63 |      "execution_count": 3,
 64 |      "metadata": {},
 65 |      "output_type": "execute_result"
 66 |     }
 67 |    ],
 68 |    "source": [
 69 |     "X_train = X_train.reshape((X_train.shape[0],X_train.shape[1],X_train.shape[2],1))\n",
 70 |     "print(X_train.shape)\n",
 71 |     "X_test = X_test.reshape((X_test.shape[0],X_test.shape[1],X_test.shape[2],1))\n",
 72 |     "print(X_test.shape)\n",
 73 |     "X_train = X_train/255\n",
 74 |     "X_test  = X_test/255\n",
 75 |     "X_train = X_train.astype('float')\n",
 76 |     "X_test  = X_test.astype('float')\n",
 77 |     "X_train[0].shape\n",
 78 |     "\n"
 79 |    ]
 80 |   },
 81 |   {
 82 |    "cell_type": "code",
 83 |    "execution_count": 4,
 84 |    "metadata": {},
 85 |    "outputs": [
 86 |     {
 87 |      "data": {
 88 |       "image/png": "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\n",
 89 |       "text/plain": [
 90 |        "<Figure size 432x288 with 1 Axes>"
 91 |       ]
 92 |      },
 93 |      "metadata": {
 94 |       "needs_background": "light"
 95 |      },
 96 |      "output_type": "display_data"
 97 |     }
 98 |    ],
 99 |    "source": [
100 |     "def show_img(img):\n",
101 |     "    img  = np.array(img,dtype='float')\n",
102 |     "    img = img.reshape((28,28))\n",
103 |     "    \n",
104 |     "    \n",
105 |     "    \n",
106 |     "    \n",
107 |     "    plt.imshow(img)\n",
108 |     "x = np.random.randint(0,X_test.shape[0])\n",
109 |     "show_img(X_train[x])"
110 |    ]
111 |   },
112 |   {
113 |    "cell_type": "code",
114 |    "execution_count": 5,
115 |    "metadata": {
116 |     "scrolled": true
117 |    },
118 |    "outputs": [
119 |     {
120 |      "name": "stdout",
121 |      "output_type": "stream",
122 |      "text": [
123 |       "WARNING:tensorflow:From C:\\Users\\Divyanshu\\Anaconda3\\envs\\sift\\lib\\site-packages\\tensorflow\\python\\framework\\op_def_library.py:263: colocate_with (from tensorflow.python.framework.ops) is deprecated and will be removed in a future version.\n",
124 |       "Instructions for updating:\n",
125 |       "Colocations handled automatically by placer.\n",
126 |       "_________________________________________________________________\n",
127 |       "Layer (type)                 Output Shape              Param #   \n",
128 |       "=================================================================\n",
129 |       "conv2d_1 (Conv2D)            (None, 28, 28, 16)        160       \n",
130 |       "_________________________________________________________________\n",
131 |       "conv2d_2 (Conv2D)            (None, 28, 28, 32)        4640      \n",
132 |       "_________________________________________________________________\n",
133 |       "conv2d_3 (Conv2D)            (None, 28, 28, 64)        18496     \n",
134 |       "_________________________________________________________________\n",
135 |       "conv2d_4 (Conv2D)            (None, 28, 28, 128)       73856     \n",
136 |       "_________________________________________________________________\n",
137 |       "global_average_pooling2d_1 ( (None, 128)               0         \n",
138 |       "_________________________________________________________________\n",
139 |       "dense_1 (Dense)              (None, 10)                1290      \n",
140 |       "=================================================================\n",
141 |       "Total params: 98,442\n",
142 |       "Trainable params: 98,442\n",
143 |       "Non-trainable params: 0\n",
144 |       "_________________________________________________________________\n",
145 |       "WARNING:tensorflow:From C:\\Users\\Divyanshu\\Anaconda3\\envs\\sift\\lib\\site-packages\\tensorflow\\python\\ops\\math_ops.py:3066: to_int32 (from tensorflow.python.ops.math_ops) is deprecated and will be removed in a future version.\n",
146 |       "Instructions for updating:\n",
147 |       "Use tf.cast instead.\n",
148 |       "Train on 54000 samples, validate on 6000 samples\n",
149 |       "Epoch 1/5\n",
150 |       "54000/54000 [==============================] - 40s 740us/step - loss: 0.7541 - acc: 0.7518 - val_loss: 0.2656 - val_acc: 0.9277\n",
151 |       "Epoch 2/5\n",
152 |       "54000/54000 [==============================] - 31s 579us/step - loss: 0.2678 - acc: 0.9204 - val_loss: 0.1491 - val_acc: 0.9577\n",
153 |       "Epoch 3/5\n",
154 |       "54000/54000 [==============================] - 31s 570us/step - loss: 0.1769 - acc: 0.9484 - val_loss: 0.1450 - val_acc: 0.9573\n",
155 |       "Epoch 4/5\n",
156 |       "54000/54000 [==============================] - 30s 558us/step - loss: 0.1405 - acc: 0.9586 - val_loss: 0.1161 - val_acc: 0.9660\n",
157 |       "Epoch 5/5\n",
158 |       "54000/54000 [==============================] - 30s 564us/step - loss: 0.1162 - acc: 0.9646 - val_loss: 0.0882 - val_acc: 0.9735\n"
159 |      ]
160 |     },
161 |     {
162 |      "data": {
163 |       "text/plain": [
164 |        "<keras.callbacks.History at 0x23c90c100f0>"
165 |       ]
166 |      },
167 |      "execution_count": 5,
168 |      "metadata": {},
169 |      "output_type": "execute_result"
170 |     }
171 |    ],
172 |    "source": [
173 |     "np.random.seed(0)\n",
174 |     "model = Sequential()\n",
175 |     "model.add(Conv2D(16,input_shape=(28,28,1),kernel_size=(3,3),activation='relu',padding='same'))\n",
176 |     "#model.add(MaxPooling2D(pool_size=(2,2)))\n",
177 |     "\n",
178 |     "model.add(Conv2D(32,kernel_size=(3,3),activation='relu',padding='same'))\n",
179 |     "#model.add(MaxPooling2D(pool_size=(2,2)))\n",
180 |     "\n",
181 |     "model.add(Conv2D(64,kernel_size=(3,3),activation='relu',padding='same'))\n",
182 |     "#model.add(MaxPooling2D())\n",
183 |     "\n",
184 |     "model.add(Conv2D(128,kernel_size=(3,3),activation='relu',padding='same'))\n",
185 |     "model.add(GlobalAveragePooling2D())\n",
186 |     "model.add(Dense(10,activation='softmax'))\n",
187 |     "\n",
188 |     "model.summary()\n",
189 |     "model.compile(loss='sparse_categorical_crossentropy',metrics=['accuracy'],optimizer='adam')\n",
190 |     "model.fit(X_train,Y_train,batch_size=32,epochs=5,validation_split=0.1,shuffle=True)"
191 |    ]
192 |   },
193 |   {
194 |    "cell_type": "code",
195 |    "execution_count": null,
196 |    "metadata": {},
197 |    "outputs": [],
198 |    "source": [
199 |     "plot_model(model,to_file='model.png')"
200 |    ]
201 |   },
202 |   {
203 |    "cell_type": "code",
204 |    "execution_count": 6,
205 |    "metadata": {},
206 |    "outputs": [],
207 |    "source": [
208 |     "model.save('Activation.h5')\n"
209 |    ]
210 |   },
211 |   {
212 |    "cell_type": "code",
213 |    "execution_count": 7,
214 |    "metadata": {},
215 |    "outputs": [
216 |     {
217 |      "name": "stdout",
218 |      "output_type": "stream",
219 |      "text": [
220 |       "10000/10000 [==============================] - 2s 198us/step\n",
221 |       "_________________________________________________________________\n",
222 |       "Layer (type)                 Output Shape              Param #   \n",
223 |       "=================================================================\n",
224 |       "conv2d_1 (Conv2D)            (None, 28, 28, 16)        160       \n",
225 |       "_________________________________________________________________\n",
226 |       "conv2d_2 (Conv2D)            (None, 28, 28, 32)        4640      \n",
227 |       "_________________________________________________________________\n",
228 |       "conv2d_3 (Conv2D)            (None, 28, 28, 64)        18496     \n",
229 |       "_________________________________________________________________\n",
230 |       "conv2d_4 (Conv2D)            (None, 28, 28, 128)       73856     \n",
231 |       "_________________________________________________________________\n",
232 |       "global_average_pooling2d_1 ( (None, 128)               0         \n",
233 |       "_________________________________________________________________\n",
234 |       "dense_1 (Dense)              (None, 10)                1290      \n",
235 |       "=================================================================\n",
236 |       "Total params: 98,442\n",
237 |       "Trainable params: 98,442\n",
238 |       "Non-trainable params: 0\n",
239 |       "_________________________________________________________________\n"
240 |      ]
241 |     }
242 |    ],
243 |    "source": [
244 |     "model.evaluate(X_test, Y_test)\n",
245 |     "model.summary()"
246 |    ]
247 |   },
248 |   {
249 |    "cell_type": "code",
250 |    "execution_count": 8,
251 |    "metadata": {},
252 |    "outputs": [
253 |     {
254 |      "data": {
255 |       "text/plain": [
256 |        "(128, 10)"
257 |       ]
258 |      },
259 |      "execution_count": 8,
260 |      "metadata": {},
261 |      "output_type": "execute_result"
262 |     }
263 |    ],
264 |    "source": [
265 |     "from keras.models import Model\n",
266 |     "import scipy as sp\n",
267 |     "gap_weights = model.layers[-1].get_weights()[0]\n",
268 |     "gap_weights.shape"
269 |    ]
270 |   },
271 |   {
272 |    "cell_type": "code",
273 |    "execution_count": 9,
274 |    "metadata": {},
275 |    "outputs": [],
276 |    "source": [
277 |     "cam_model  = Model(inputs=model.input,outputs=(model.layers[-3].output,model.layers[-1].output))"
278 |    ]
279 |   },
280 |   {
281 |    "cell_type": "code",
282 |    "execution_count": 10,
283 |    "metadata": {},
284 |    "outputs": [
285 |     {
286 |      "name": "stdout",
287 |      "output_type": "stream",
288 |      "text": [
289 |       "_________________________________________________________________\n",
290 |       "Layer (type)                 Output Shape              Param #   \n",
291 |       "=================================================================\n",
292 |       "conv2d_1_input (InputLayer)  (None, 28, 28, 1)         0         \n",
293 |       "_________________________________________________________________\n",
294 |       "conv2d_1 (Conv2D)            (None, 28, 28, 16)        160       \n",
295 |       "_________________________________________________________________\n",
296 |       "conv2d_2 (Conv2D)            (None, 28, 28, 32)        4640      \n",
297 |       "_________________________________________________________________\n",
298 |       "conv2d_3 (Conv2D)            (None, 28, 28, 64)        18496     \n",
299 |       "_________________________________________________________________\n",
300 |       "conv2d_4 (Conv2D)            (None, 28, 28, 128)       73856     \n",
301 |       "_________________________________________________________________\n",
302 |       "global_average_pooling2d_1 ( (None, 128)               0         \n",
303 |       "_________________________________________________________________\n",
304 |       "dense_1 (Dense)              (None, 10)                1290      \n",
305 |       "=================================================================\n",
306 |       "Total params: 98,442\n",
307 |       "Trainable params: 98,442\n",
308 |       "Non-trainable params: 0\n",
309 |       "_________________________________________________________________\n"
310 |      ]
311 |     }
312 |    ],
313 |    "source": [
314 |     "cam_model.summary()"
315 |    ]
316 |   },
317 |   {
318 |    "cell_type": "code",
319 |    "execution_count": 11,
320 |    "metadata": {},
321 |    "outputs": [
322 |     {
323 |      "data": {
324 |       "text/plain": [
325 |        "(10000, 28, 28, 128)"
326 |       ]
327 |      },
328 |      "execution_count": 11,
329 |      "metadata": {},
330 |      "output_type": "execute_result"
331 |     }
332 |    ],
333 |    "source": [
334 |     "features,results = cam_model.predict(X_test)\n",
335 |     "features.shape"
336 |    ]
337 |   },
338 |   {
339 |    "cell_type": "code",
340 |    "execution_count": 15,
341 |    "metadata": {},
342 |    "outputs": [
343 |     {
344 |      "data": {
345 |       "image/png": "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\n",
346 |       "text/plain": [
347 |        "<Figure size 432x288 with 1 Axes>"
348 |       ]
349 |      },
350 |      "metadata": {},
351 |      "output_type": "display_data"
352 |     },
353 |     {
354 |      "data": {
355 |       "image/png": "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\n",
356 |       "text/plain": [
357 |        "<Figure size 432x288 with 1 Axes>"
358 |       ]
359 |      },
360 |      "metadata": {},
361 |      "output_type": "display_data"
362 |     },
363 |     {
364 |      "data": {
365 |       "image/png": "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\n",
366 |       "text/plain": [
367 |        "<Figure size 432x288 with 1 Axes>"
368 |       ]
369 |      },
370 |      "metadata": {},
371 |      "output_type": "display_data"
372 |     },
373 |     {
374 |      "data": {
375 |       "image/png": "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\n",
376 |       "text/plain": [
377 |        "<Figure size 432x288 with 1 Axes>"
378 |       ]
379 |      },
380 |      "metadata": {},
381 |      "output_type": "display_data"
382 |     },
383 |     {
384 |      "data": {
385 |       "image/png": "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\n",
386 |       "text/plain": [
387 |        "<Figure size 432x288 with 1 Axes>"
388 |       ]
389 |      },
390 |      "metadata": {},
391 |      "output_type": "display_data"
392 |     },
393 |     {
394 |      "data": {
395 |       "image/png": "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\n",
396 |       "text/plain": [
397 |        "<Figure size 432x288 with 1 Axes>"
398 |       ]
399 |      },
400 |      "metadata": {},
401 |      "output_type": "display_data"
402 |     },
403 |     {
404 |      "data": {
405 |       "image/png": "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\n",
406 |       "text/plain": [
407 |        "<Figure size 432x288 with 1 Axes>"
408 |       ]
409 |      },
410 |      "metadata": {},
411 |      "output_type": "display_data"
412 |     },
413 |     {
414 |      "data": {
415 |       "image/png": "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\n",
416 |       "text/plain": [
417 |        "<Figure size 432x288 with 1 Axes>"
418 |       ]
419 |      },
420 |      "metadata": {},
421 |      "output_type": "display_data"
422 |     },
423 |     {
424 |      "data": {
425 |       "image/png": "iVBORw0KGgoAAAANSUhEUgAAAP8AAAEGCAYAAACq4kOvAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjAsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+17YcXAAAgAElEQVR4nO3df1SUVf4H8PcAIii/RMQQVERREQYRB9FNMX8gpRuumqbZScNks5+rraXnbEnlrp6wPeXmulGmbHXQtfJHaS6bippphDJaUkYkpUgogvxGGLjfPzzOV4LnXgQGsOf9OqcTzn2eee48M595ZuZzP/cahBACRKQ7dh3dASLqGAx+Ip1i8BPpFIOfSKcY/EQ65dCeB3N2d4frHb3b85B0W7NdIspgs3u+Tt1z2/Sg7JcCVJWUNGvbVgX/vn378PTTT6Ourg6PPPIIVqxYId3e9Y7emPXmP1pzSNIRO9S3ql3GYMM3FgAQiuC22Oi6++Efn2z2ti3+2F9XV4fHH38cn376KbKyspCSkoKsrKyW3h0RtbMWB396ejoGDRqEgIAAODo6Yu7cudi1a1db9o2IbKjFwZ+Xl4e+ffta/+3n54e8vLxG2yUlJcFkMsFkMjX7uwgR2V6Lg7+pUcEGQ+PvOfHx8cjIyEBGRgac3d1bejgiamMtDn4/Pz+cP3/e+u8LFy6gT58+bdIpIrK9Fgd/REQEsrOzce7cOdTU1GDr1q2IjY1ty74RkQ21ON/g4OCAN954AzExMairq0NcXByCg4Pbsm/UCajSaV1Q2+J2J1yT7uuEamm7PSzSdrtWpPMssJe216CrtL0SztL2ajhJ7ttRum+95Jp9K4+4VcnGqVOnYurUqa25CyLqIBzeS6RTDH4inWLwE+kUg59Ipxj8RDrF4CfSqXat5yfbkOXiW5OHB9S5+O4oV+yvnavvosjTGxRjDOoUuXghubap8vj1ipJcVXuVIs9fiW6abdX12mMAAMDerk7Sr+Zfz3nlJ9IpBj+RTjH4iXSKwU+kUwx+Ip1i8BPpFFN9twEHRUpM1t4VNdJ9HRWpvG6olLZ3R4W0XZVKlKlr5ctTlY6TUZUDq86riyIFek1SElxb00W6b72DpKRXNP8x88pPpFMMfiKdYvAT6RSDn0inGPxEOsXgJ9IpBj+RTjHP3w5aO/21oyKnbA/tEs/WlvSqVKB7i/cVjvLS1TpnebtqJVzpFNdV8sc9pN930vbA0xek7Q5d5M/52S7aS9V/6zRMum8BvLWPa9B+Lfwar/xEOsXgJ9IpBj+RTjH4iXSKwU+kUwx+Ip1i8BPpFPP8zSSrmZfl2QF1Ll1Vr68aJyCrey+Dq3RfVa5cxdLbTb6Bp/Z8AQEuBdJdfR2KpO2qqbtlNfmOFvmxp+09Jm0/lim/bg6StgIGL+16fxEuf05ky3/bKV6LN2tV8Pv7+8PV1RX29vZwcHBARkZGa+6OiNpRq6/8Bw8ehJeXV1v0hYjaEb/zE+lUq4LfYDBgypQpGDlyJJKSkprcJikpCSaTCSaTCVUlJa05HBG1oVZ97D969Cj69OmDS5cuITo6GkOHDkVUVFSDbeLj4xEfHw8A8B4yuDWHI6I21Korf58+fQAA3t7emDFjBtLT09ukU0Rkey0O/oqKCpSVlVn/Tk1NRUhISJt1jIhsq8Uf+wsKCjBjxgwAgMViwQMPPIC77767zTrW3roq5q+X5eoNijneVeMAZEtJA0A1HBXt2nXvTi7y+7b3lc/LP6zHJWm7xVAmbXeFdrvqvNQoHneNZO57FRcH7SWyAcDBTn7e5CMvAPkIBQCF2k09FHt3Q5Vmm2pMyM1aHPwBAQE4depUS3cnog7GVB+RTjH4iXSKwU+kUwx+Ip1i8BPplG5KelWpPCdFu0GSQlGlV1TLPVcrngZZKg+Qp8wcnOTHHu4pyTkBuIqe0nZVObLsvMrSgIB6iW3VEt4FFu2lrguq5f0299OeHhsA3OUze6NWvso2HHppp+uuood0X/l54RLdRKTA4CfSKQY/kU4x+Il0isFPpFMMfiKdYvAT6dRvJs/f2mWuW7NUtYNq367yPH1lX8X016Xy0tXqS9qPrUu9PM9/RWhPAw0AlUK+v7/hovz+K7T79lOFPJ9dUSZvry1VLNFdp913D4t8fEO5Il/ucYe8FBoe8ubcnr6abfnwke4rK2W+lanYeeUn0ikGP5FOMfiJdIrBT6RTDH4inWLwE+kUg59Ip26rPL+sdlxVr99VkeeX1esD8pp8g4N8qWhXozwnHNBF3r6nLELaXvOjdm63Blek+3b9Wj7GoKpCnuev6tZP2u5aqT31dzm6S/ctlSxFDQCVkE+/LVvCuwjyx62aQyFo4LfSdtXrsUgyT0IJ3KX71krClnl+IlJi8BPpFIOfSKcY/EQ6xeAn0ikGP5FOMfiJdOq2yvNL56dXzB+vyuOrCDvtnHFNsLzevmcXed/e/+EuaXu3g9XSdlzVPi+lHp7SXd1QKm2vhXwC+rpK+RiHCxii2VZaLc9nW67K8/y4Km+GbPiEfPVvXPAeIG3P8h4mbe8mPbh8fgnVOhCy50S11kHD4yjExcXB29sbISEh1tuKiooQHR2NwMBAREdHo7i4uNkHJKLOQRn8CxcuxL59+xrctnbtWkyaNAnZ2dmYNGkS1q5da7MOEpFtKIM/KioKnp4NPzru2rULCxYsAAAsWLAAO3futE3viMhmWvSdv6CgAD4+1+cZ8/HxwaVL2uO3k5KSkJSUBACoKilpyeGIyAZs/mt/fHw8MjIykJGRAWd3+Q88RNR+WhT8vXv3Rn5+PgAgPz8f3t7yFU2JqPNpUfDHxsYiOTkZAJCcnIzp06e3aaeIyPaU3/nnzZuHtLQ0FBYWws/PDy+++CJWrFiBOXPmYNOmTejXrx+2b9/eJp1R5TfleX7ttuv3La9Lh4P8VFQO0M459+wmz0ebX9Geox0Auh05Im2Ha7i8fYJ2k4/neemufSFvl80RDwBXFbXnspr6Lo7y9Q4sTopkvKN8jAFkwyPk0zsoxxCUQ76mQKWLfK4Bp27anVPNBSBbY+JW6vmVwZ+SktLk7fv372/2QYio8+HwXiKdYvAT6RSDn0inGPxEOsXgJ9KpTlXS25pUX2vV9pZP1fw7r5812/534HfSffOPnVYc3FXeLq8ehUuIdlWlqmRXtTS5K8qk7T1QJG0XkutLrZ385VftIX9OKjxcpO2yqcFVU3Pfrpqf6OOVn0i3GPxEOsXgJ9IpBj+RTjH4iXSKwU+kUwx+Ip3qVHl+g6rsVkI9ZbG8/FO4yafXvgoPzbYL5fLyTvtekdJ23CtvHuQlXw5alrNW5bPLIB9joJy6W3FeZftXQJ6nr1aUE19TtNeo5ueWUI0pseWYk/bCKz+RTjH4iXSKwU+kUwx+Ip1i8BPpFIOfSKcY/EQ61any/Kpph2U5ZVXetYtiruYAN/lczedKhmu2hfgVSPc9+Uc/afuQbtnS9q5X5X2X1eS7K+r5HRT1/Kop0VXjK2RTptdJ59YGahUvT9UYg0qLfPpsma4O8umznVHV4vvuLHjlJ9IpBj+RTjH4iXSKwU+kUwx+Ip1i8BPpFIOfSKc6VZ6/XvFeJGtX5aNV89M72ilywjXaOWWHQPlcAHe5yuvxXRS5+FPFPtJ2tzLtfHlhV+256wHAqUp+Xlwq5Plsg2KtBUt37fPWVbEMtn2VfAPVXANXqntqtgmLYnlv+VQDcHKQj1FozdwU7UV55Y+Li4O3tzdCQkKstyUkJMDX1xdhYWEICwvD3r17bdpJImp7yuBfuHAh9u3b1+j2pUuXwmw2w2w2Y+rUqTbpHBHZjjL4o6Ki4Onp2R59IaJ21OIf/N544w2EhoYiLi4OxcXaa8UlJSXBZDLBZDKhqqSkpYcjojbWouBfsmQJcnJyYDab4ePjg2eeeUZz2/j4eGRkZCAjIwPO7u4t7igRta0WBX/v3r1hb28POzs7LF68GOnp6W3dLyKysRYFf35+vvXvHTt2NMgEENHtQZnnnzdvHtLS0lBYWAg/Pz+8+OKLSEtLg9lshsFggL+/P95888026Ywqzy+r3y5TJGZV7blX5F9JQnvla7ap5hJwgHwcgEXxNIT3uCxt79JDO1fvqJjHwBHyunUnRbu94rEJyXNqqZePIThbKR/f0O1r+f6OTtqP/Vq5s3RfUX8rK93fnpTBn5KS0ui2RYsW2aQzRNR+OLyXSKcY/EQ6xeAn0ikGP5FOMfiJdKpTlfSqpu6WlXDaKUpLVSW9Ltny8tGfL2uX/PoEylNxqlSfk718KelaxVLUMq1ZQrs5PBTTb8ueF0877WHhAGDnIi+z/iRgoLS964/aaUp7j9alZ2+Hkl0VXvmJdIrBT6RTDH4inWLwE+kUg59Ipxj8RDrF4CfSqU6V51eV9NZAOx9eWy/PVzvZyfPR3UWFtB3F2vtfSpfn4Q2KXHphr/7S9gqDvPzUrW+ZZttAR+02QH3OVaXQJfCQtvfDz5pt1XCS7lsFeZ7f2UX+nLpIltFWjQuRvdaaQ1Wm3Rnwyk+kUwx+Ip1i8BPpFIOfSKcY/EQ6xeAn0ikGP5FOdapkpCr3KstJV1cqcsaVbtL2Yid5/ba7m3a9f1fF9Naq6bMNl8ul7ara8nIX7XEGPXvnSvf9sayftD0tfYi03clePgeDZcgIzbZYn5PSfSshH99Qr5j/wR3aYzdUS4urXouqeRCY5yeiTovBT6RTDH4inWLwE+kUg59Ipxj8RDrF4CfSqU6VjFTNhS5bTrrGUV5/XXNVUZ99QX4qSmq8tBtVZ1ExVYBian0oSuYhJPX+eR6+8kOb5edlVOkv0vZvr/aQtr901wbNtt64JN33A9wnbS+t6S5td0WRZltrX/iqeRBuB8pHcP78eUyYMAFBQUEIDg7G66+/DgAoKipCdHQ0AgMDER0djeJi+QIMRNS5KIPfwcEBr776Kr799lscP34cGzZsQFZWFtauXYtJkyYhOzsbkyZNwtq1a9ujv0TURpTB7+Pjg/DwcACAq6srgoKCkJeXh127dmHBggUAgAULFmDnzp227SkRtalb+uqTm5uLzMxMREZGoqCgAD4+PgCuv0FcutT097ekpCQkJSUBAKpKSlrZXSJqK83+1aK8vByzZs3Ca6+9Bjc3eZHMzeLj45GRkYGMjAw4u7u3qJNE1PaaFfy1tbWYNWsW5s+fj5kzZwIAevfujfz8fABAfn4+vL29bddLImpzyo/9QggsWrQIQUFBWLZsmfX22NhYJCcnY8WKFUhOTsb06dNt2lEA6CIpbfVwlC+xXe0nn+a51E3xqeQXyak6L98VuYp2edeVqT5DtXbJcDrk04KH18jTbaq+BQ2WZ3mG/ZKt2Say5andinD510SvH+Rl3LLlyVVTc6uWNlctJ387UAb/0aNH8e6778JoNCIsLAwA8Le//Q0rVqzAnDlzsGnTJvTr1w/bt2+3eWeJqO0og3/s2LEQoul36P3797d5h4iofdz+w5SIqEUY/EQ6xeAn0ikGP5FOMfiJdKpTlfS2hqoc2FmyXDMAOLvJ24Wbdl732mD5Et1VNfIpqGuuyktTIR+iAHSr02xyLJFPQf31HT2l7f27aZfFAsCkMZ9L2wu/0J7i+lyufGCYl0O+tL1IsXy4kFzbhGJqblWeXxclvUT028TgJ9IpBj+RTjH4iXSKwU+kUwx+Ip1i8BPp1G8mz29rsnEEqiW6FaXjMHjKxyjUWeQ55+5O2nODd0Gt/L4r5fdd4isfo3CyWJ6rv6OL9mQEjn6l0n1VS3A7KQZAyGr2a9HyuQB+K3jlJ9IpBj+RTjH4iXSKwU+kUwx+Ip1i8BPpFIOfSKeY528DqrkEVLn2GjvF8uIWeXtpufYKSl0c5cd2cSyTtjvmyefOr1Tkyw9jlGabk5M8T38F8rkGyhX1/LXQnktA1gb8Nur1VX77j5CImsTgJ9IpBj+RTjH4iXSKwU+kUwx+Ip1i8BPplDLPf/78eTz00EP45ZdfYGdnh/j4eDz99NNISEjAW2+9hV69egG4vmz31KlTbd7h25GdYo54R7saafs1O/m6APWl2u01lfIxAkVu8jUDHFzk6xlUOcnr/UugPQbBARbpvmRbyuB3cHDAq6++ivDwcJSVlWHkyJGIjo4GACxduhR//vOfbd5JImp7yuD38fGBj48PAMDV1RVBQUHIy8uzeceIyLZu6Tt/bm4uMjMzERkZCQB44403EBoairi4OBQXFze5T1JSEkwmE0wmE6pK5ENFiaj9NDv4y8vLMWvWLLz22mtwc3PDkiVLkJOTA7PZDB8fHzzzzDNN7hcfH4+MjAxkZGTA2d29zTpORK3TrOCvra3FrFmzMH/+fMycORMA0Lt3b9jb28POzg6LFy9Genq6TTtKRG1LGfxCCCxatAhBQUFYtmyZ9fb8/P9fQXXHjh0ICQmxTQ+JyCaUP/gdPXoU7777LoxGI8LCwgBcT+ulpKTAbDbDYDDA398fb775ps07e7tSlfyqpv627yZPiVU4ape2XiuSl71CPns2LKXyVF6pi7zd4q39O48r5OXEqlJoVQpVNv22Hkp2VZTBP3bsWAjR+MXLnD7R7Y1vf0Q6xeAn0ikGP5FOMfiJdIrBT6RTDH4ineLU3Z2AeupveZ7fw+GqZlutd7l0X4u3fAprFXvUSdsdoV2urMrjC8US3dcgL3UmOV75iXSKwU+kUwx+Ip1i8BPpFIOfSKcY/EQ6xeAn0imDaKpe10a8vLzg7+9v/ffly5etU393Np21b521XwD71lJt2bfc3FwUFhY2a9t2Df5fM5lMyMjI6KjDS3XWvnXWfgHsW0t1VN/4sZ9Ipxj8RDpln5CQkNCRHRg5cmRHHl6qs/ats/YLYN9aqiP61qHf+Ymo4/BjP5FOMfiJdKpDgn/fvn0YMmQIBg0ahLVr13ZEFzT5+/tb1ygwmUwd2pe4uDh4e3s3WBClqKgI0dHRCAwMRHR0tOYaiR3Rt4SEBPj6+iIsLAxhYWHYu3dvh/Tt/PnzmDBhAoKCghAcHIzXX38dQMefO61+ddh5E+3MYrGIgIAAkZOTI65duyZCQ0PFmTNn2rsbmvr37y8uX77c0d0QQghx6NAhceLECREcHGy9bfny5WLNmjVCCCHWrFkjnn322U7Tt1WrVonExMQO6c/NLl68KE6cOCGEEKK0tFQEBgaKM2fOdPi50+pXR523dr/yp6enY9CgQQgICICjoyPmzp2LXbt2tXc3bgtRUVHw9PRscNuuXbuwYMECAMCCBQuwc+fOjuhak33rLHx8fBAeHg6g4bLyHX3utPrVUdo9+PPy8tC3b1/rv/38/Dr0BPyawWDAlClTMHLkSCQlJXV0dxopKCiAj48PgOsvpkuXLnVwjxpqzrLt7enmZeU707lryXL3ba3dg180kVk0GORztbWno0eP4uTJk/j000+xYcMGHD58uKO7dNto7rLt7eXXy8p3Fi1d7r6ttXvw+/n54fz589Z/X7hwAX369Gnvbmi60Rdvb2/MmDGj0y093rt3b+sKyfn5+fD29u7gHv2/zrRsu9ay8h197jrTcvftHvwRERHIzs7GuXPnUFNTg61btyI2Nra9u9GkiooKlJWVWf9OTU3tdEuPx8bGIjk5GQCQnJyM6dOnd3CP/l9nWbZdaCwr39HnTqtfHXbe2v0nRiHEnj17RGBgoAgICBCrV6/uiC40KScnR4SGhorQ0FAxbNiwDu/b3LlzxR133CEcHByEr6+vePvtt0VhYaGYOHGiGDRokJg4caK4cuVKp+nbgw8+KEJCQoTRaBT33nuvuHjxYof07ciRIwKAMBqNYvjw4WL48OFiz549HX7utPrVUeeNw3uJdIoj/Ih0isFPpFMMfiKdYvAT6RSDn0inbB789vb2CAsLQ0hICGbPno3KysoW31daWhp+//vfAwB2794trQi8evUq/vnPf97yMRISErBu3bom2/79738jJCQEwcHBGDZsmHW7hQsX4oMPPrjlY7W1LVu2oFevXtbqsLffflu5z40qxuHDh2PKlCn45ZdfbumY/v7+zZ4tFtA+vxcvXsR9990HQPt53rlzJ7Kysm6pf7cqOTkZgYGBCAwMtI4JaMo//vEPDBkyBMHBwXj22Wett58+fRpjxoxBcHAwjEYjqqurAQApKSkwGo0IDQ3F3XffbT1np06dwpgxY2A0GnHvvfeitLQUAPD+++9bn8ewsDDY2dnBbDajrKyswe1eXl7405/+BABYunSp9fbBgwfDw8ND/mBtnUvs3r279e8HHnhAvPrqqw3a6+vrRV1dXbPu6+DBg2LatGnN2vbcuXMNKs6aS6vCau/evWLEiBEiLy9PCCFEVVWVSEpKEkIIsWDBArF9+/ZbPlZb27x5s3j88cdvaZ+bqxhXrlwpnnzyyUbbWCyWZu3fHM2pYNN6nm19nq9cuSIGDBggrly5IoqKisSAAQNEUVFRo+0OHDggJk2aJKqrq4UQQhQUFAghhKitrRVGo1GYzWYhhBCFhYXCYrGI2tpa0atXL+t5Wr58uVi1apUQQgiTySTS0tKEEEJs2rRJ/OUvf2l0vNOnT4sBAwY02efw8HBx6NChRrevX79ePPzww9LH264f+8eNG4cffvgBubm5CAoKwmOPPYbw8HCcP38eqampGDNmDMLDwzF79myUl19fV37fvn0YOnQoxo4di48++sh6X1u2bMETTzwB4Hqxy4wZMzB8+HAMHz4cX3zxBVasWIGcnByEhYVh+fLlAIDExEREREQgNDQUq1atst7XX//6VwwZMgSTJ0/G2bNnm+z7mjVrsG7dOuvwXycnJyxevLjRdi+99BIiIiIQEhKC+Ph4ay3D+vXrMWzYMISGhmLu3LkAgEOHDlnfqUeMGGEdXdhRoqKi8MMPPwAAXFxc8MILLyAyMhLHjh3D/v37MWLECBiNRsTFxeHatWvW/RITEzFq1CiMGjXKuv/HH3+MyMhIjBgxApMnT0ZBQYF1+1OnTmHixIkIDAzEW2+9BeB6oUtTI9tuPM9ffPEFdu/ejeXLlyMsLAw5OTnWCjkAyM7ObvU8eP/9738RHR0NT09P9OjRA9HR0di3b1+j7TZu3IgVK1aga9euAGAdJpyamorQ0FAMHz4cANCzZ0/Y29tDCAEhBCoqKiCEQGlpqfV1dPbsWURFRQEAoqOj8eGHHzY6XkpKCubNm9fo9uzsbFy6dAnjxo1r9j43a7fgt1gs+PTTT2E0GgFcf9APPfQQMjMz0b17d6xevRqfffYZTp48CZPJhL///e+orq7G4sWL8fHHH+PIkSOaH0mfeuopjB8/HqdOncLJkycRHByMtWvXYuDAgTCbzUhMTERqaiqys7ORnp4Os9mMEydO4PDhwzhx4gS2bt2KzMxMfPTRR/jqq6+aPMY333zTrBfXE088ga+++grffPMNqqqq8MknnwAA1q5di8zMTJw+fRr/+te/AADr1q3Dhg0bYDabceTIETg7Oze6v3HjxjX4mHfjv88++6zJ43/44YcIDQ3Ffffd16CGojk++eQT6/NTUVGBkJAQfPnllzCZTFi4cCG2bduGr7/+GhaLBRs3brTu5+bmhvT0dDzxxBPWj6Bjx47F8ePHkZmZiblz5+KVV16xbn/69Gns2bMHx44dw0svvYSLFy8q+/a73/0OsbGxSExMhNlsxsCBA+Hu7g6z2QwA2Lx5MxYuXNhov8TExCbP31NPPdVo2+ZWnH7//fc4cuQIIiMjMX78eOtr5vvvv4fBYEBMTAzCw8Otj7lLly7YuHEjjEYj+vTpg6ysLCxatAgAEBISgt27dwMAtm/f3uRztm3btiYDOSUlBffff3+jwriffvoJ586dw8SJE5s8lzfYPPirqqqss+L069fP+qD79++P0aNHAwCOHz+OrKws3HnnnQgLC0NycjJ++uknfPfddxgwYAACAwNhMBjw4IMPNnmMAwcOYMmSJQCu/8bg7u7eaJvU1FSkpqZixIgRCA8Px3fffYfs7GwcOXIEM2bMQLdu3eDm5tbqOoODBw8iMjISRqMRBw4cwJkzZwAAoaGhmD9/Pt577z04ODgAAO68804sW7YM69evx9WrV6233+zIkSMwm82N/ps8eXKjbe+9917k5ubi9OnTmDx5srV2XWXChAkICwtDaWkpVq5cCeD6eZw1axaA62/UAwYMwODBgwFcr4W/udrxxgtz3rx5OHbsGIDrBVsxMTEwGo1ITEy0ngcAmD59OpydneHl5YUJEya0uJDlkUcewebNm1FXV4dt27bhgQceaLTN8uXLmzx/69evb7StaGbFqcViQXFxMY4fP47ExETMmTMHQghYLBZ8/vnneP/99/H5559jx44d2L9/P2pra7Fx40ZkZmbi4sWLCA0NxZo1awAA77zzDjZs2ICRI0eirKwMjo6ODY715Zdfolu3bk1+Ktq6dWuTbwpbt27FfffdB3t7e+2TB6Dxq62NOTs7W9+db9a9e3fr30IIREdHIyUlpcE2ZrO5zcp9hRBYuXIl/vjHPza4/bXXXmvWMYKDg3HixAnpu2l1dTUee+wxZGRkoG/fvkhISLD+4LNnzx4cPnwYu3fvxssvv4wzZ85gxYoVmDZtGvbu3YvRo0fjs88+w9ChQxvc57hx45r8OrBu3bpGbwA9e/a0/r148WI899xzyscFXH/D8vLyanCbk5OT9cXTVFDc7Obzd+PvJ598EsuWLUNsbCzS0tJw8wzxvz7fLX2OZ82ahRdffBETJ07EyJEjGzz+GxITE/H+++83uj0qKqrRG4Cfnx/S0tKs/75w4QLuuuuuRvv6+flh5syZMBgMGDVqFOzs7FBYWAg/Pz+MHz/eei6nTp2KkydPWsuJBw4cCACYM2eO9UfMoUOHIjU1FcD1Tw579uxpcCytAD916hQsFkuTn0a3bt2KDRs2NLr91zpFqm/06NE4evSo9ftiZWUlvv/+ewwdOhTnzp1DTk4OADR6c7hh0qRJ1o+hdXV1KC0thaura4OgiYmJwTvvvGP9LSEvLw+XLl1CVFQUduzYgaqqKpSVleHjjz9u8hgrV67Es88+a/3qce3atUYvnhuB7uXlhfLyct3a/HcAAAMxSURBVGsGoL6+3jp/2yuvvIKrV6+ivLwcOTk5MBqNeO6552AymfDdd981Ou6tXPlvrg7bvXs3goKCrP/+9ZvKrRg6dChyc3Otz8+7776L8ePHW9u3bdtm/f+YMWMAACUlJfD19QWARr+a79q1C9XV1bhy5QrS0tIQERHRrH78+jl1cnJCTEwMlixZgocffrjJfW7lyh8TE4PU1FQUFxejuLgYqampiImJabTdH/7wBxw4cADA9YCtqamBl5cXYmJicPr0aVRWVsJiseDQoUMYNmwYfH19kZWVhcuXLwMA/ve//1mfmxsTitTX12P16tV49NFHrcepr6/H9u3brb8R3UzrO/3Zs2dRXFxsfR5kbH7lb45evXphy5YtmDdvnvWHpNWrV2Pw4MFISkrCtGnT4OXlhbFjx+Kbb75ptP/rr7+O+Ph4bNq0Cfb29ti4cSPGjBmDO++8EyEhIbjnnnuQmJiIb7/91npSXFxc8N577yE8PBz3338/wsLC0L9//yZ/PAGuv4sXFBRg8uTJEELAYDAgLi6uwTYeHh5YvHgxjEYj/P39rS/quro6PPjggygpKYEQAkuXLoWHhweef/55HDx4EPb29hg2bBjuueeeVp3H9evXY/fu3XBwcICnpye2bNkCACgsLFRevWWcnJywefNmzJ49GxaLBREREQ1epNeuXUNkZCTq6+utb9AJCQmYPXs2fH19MXr0aJw7d866/ahRozBt2jT8/PPPeP7559GnTx/k5uYq+zF37lwsXrwY69evxwcffICBAwdi/vz5+OijjzBlypQWP74bPD098fzzz1uftxdeeME6VdkjjzyCRx99FCaTCXFxcYiLi0NISAgcHR2RnJwMg8GAHj16YNmyZYiIiIDBYMDUqVMxbdo0AMCqVasQFRWFLl26oH///tbnJiUlxXqVnjlzZoM3scOHD8PPzw8BAQGN+vqf//ynyYk+U1JSMHfu3GZ9mmJVnw588skn+PHHH5v8ket2t27dOpSUlODll1/u6K7cdhj8dNuaMWMGcnJycODAgUa/WZAag59IpzrFD35E1P4Y/EQ6xeAn0ikGP5FOMfiJdOr/ABiVk4dX2oqSAAAAAElFTkSuQmCC\n",
426 |       "text/plain": [
427 |        "<Figure size 432x288 with 1 Axes>"
428 |       ]
429 |      },
430 |      "metadata": {},
431 |      "output_type": "display_data"
432 |     },
433 |     {
434 |      "data": {
435 |       "image/png": "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\n",
436 |       "text/plain": [
437 |        "<Figure size 432x288 with 1 Axes>"
438 |       ]
439 |      },
440 |      "metadata": {},
441 |      "output_type": "display_data"
442 |     }
443 |    ],
444 |    "source": [
445 |     "for idx in range(10):\n",
446 |     "    features_for_one_img = features[idx,:,:,:]\n",
447 |     "    #height_roomout = X_train.shape[1]/features_for_one_img.shape[0]\n",
448 |     "    #width_roomout  = X_train.shape[2]/features_for_one_img.shape[1]\n",
449 |     "    #print(height_roomout,width_roomout)\n",
450 |     " \n",
451 |     "    #cam_features = sp.ndimage.zoom(features_for_one_img, (height_roomout, width_roomout, 1), order=2)\n",
452 |     "    #print(cam_features.shape)\n",
453 |     "    pred = np.argmax(results[idx])\n",
454 |     "    cam_features = features_for_one_img\n",
455 |     "    \n",
456 |     "    \n",
457 |     "    plt.figure(facecolor='white')\n",
458 |     "    cam_weights = gap_weights[:,pred]\n",
459 |     "    cam_output  = np.dot(cam_features,cam_weights)\n",
460 |     "    #print(features_for_one_img.shape)\n",
461 |     "\n",
462 |     "    buf = 'Predicted Class = ' +str( pred )+ ', Probability = ' + str(results[idx][pred])\n",
463 |     "\n",
464 |     "    plt.xlabel(buf)\n",
465 |     "\n",
466 |     "    plt.imshow(np.squeeze(X_test[idx],-1), alpha=0.5)\n",
467 |     "\n",
468 |     "    plt.imshow(cam_output, cmap='jet', alpha=0.5)\n",
469 |     "\n",
470 |     "     \n",
471 |     "\n",
472 |     "    plt.show()"
473 |    ]
474 |   }
475 |  ],
476 |  "metadata": {
477 |   "kernelspec": {
478 |    "display_name": "Python 3",
479 |    "language": "python",
480 |    "name": "python3"
481 |   },
482 |   "language_info": {
483 |    "codemirror_mode": {
484 |     "name": "ipython",
485 |     "version": 3
486 |    },
487 |    "file_extension": ".py",
488 |    "mimetype": "text/x-python",
489 |    "name": "python",
490 |    "nbconvert_exporter": "python",
491 |    "pygments_lexer": "ipython3",
492 |    "version": "3.6.9"
493 |   }
494 |  },
495 |  "nbformat": 4,
496 |  "nbformat_minor": 2
497 | }
498 | 


--------------------------------------------------------------------------------