Add colonoscopy detection in a video

6-WindowsPolypsDetection.ipynb

@@ -257,6 +257,64 @@
     "plt.show()"
    ]
   },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import numpy as np\n",
+    "import cv2\n",
+    "\n",
+    "font = cv2.FONT_HERSHEY_SIMPLEX \n",
+    "\n",
+    "cap= cv2.VideoCapture(r'C:\\Users\\HP\\Desktop\\polyps1.mp4')\n",
+    "\n",
+    "#For real time camera\n",
+    "#cap = cv2.VideoCapture(0)\n",
+    "\n",
+    "while(True):\n",
+    "    # Capture frame-by-frame\n",
+    "    ret, frame = cap.read()\n",
+    "    width  = int(cap.get(3)) \n",
+    "    height = int(cap.get(4)) \n",
+    "    max_pred = 0.0 # maximum prediction\n",
+    "    max_box = []   # box for the polyp detection\n",
+    "    # Our operations on the frame come here: find the box with max proba for being polyps\n",
+    "    for win_size in window_sizes:\n",
+    "    # Loop on both dimensions of the image\n",
+    "        for top in range(0, height - win_size + 1, step):\n",
+    "            for left in range(0, width - win_size + 1, step):\n",
+    "                # compute the (top, left, bottom, right) of the bounding box\n",
+    "                box = (top, left, top + win_size, left + win_size)\n",
+    "\n",
+    "                # crop the original image\n",
+    "                cropped_img = frame[box[0]:box[2], box[1]:box[3],:]\n",
+    "\n",
+    "                # normalize the cropped image (the same processing used for the CNN dataset)\n",
+    "                cropped_img = cropped_img * 1./255\n",
+    "                # reshape from (150, 150, 3) to (1, 150, 150, 3) for prediction\n",
+    "                cropped_img = cropped_img.reshape((1, cropped_img.shape[0], cropped_img.shape[1], cropped_img.shape[2]))\n",
+    "\n",
+    "                # make a prediction for only one cropped small image \n",
+    "                preds = model.predict(cropped_img, batch_size=None, verbose=0)\n",
+    "                # print(box[0],box[2],box[1],box[3], preds[0][0])\n",
+    "                if preds[0][0]> max_pred:\n",
+    "                    max_pred = preds[0][0]\n",
+    "                    max_box = box\n",
+    "                    \n",
+    "    frame = cv2.rectangle(frame, (max_box[0],max_box[1]),(max_box[2],max_box[3]) , (255, 0, 0) , 2) \n",
+    "    cv2.putText(frame,'Probability: '+str(max_pred) , (50, 50),  font, 1,  (0, 255, 255),  2, cv2.LINE_4)\n",
+    "    # Display the resulting frame\n",
+    "    cv2.imshow('frame',frame)\n",
+    "    if cv2.waitKey(1) & 0xFF == ord('q'):\n",
+    "        break\n",
+    "\n",
+    "# When everything done, release the capture\n",
+    "cap.release()\n",
+    "cv2.destroyAllWindows()"
+   ]
+  },
   {
    "cell_type": "markdown",
    "metadata": {},
@@ -301,7 +359,7 @@
    "name": "python",
    "nbconvert_exporter": "python",
    "pygments_lexer": "ipython3",
-   "version": "3.5.2"
+   "version": "3.6.8"
   }
  },
  "nbformat": 4,