scratchpad

patches_from_gather_nd.ipynb

{
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/gist/ariG23498/61dd96b5ec1460ac837d0fced4feab5c/scratchpad.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "from google.colab import drive\n",
        "drive.mount('/content/drive')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "zcDnON6hMBDT",
        "outputId": "a5695c1b-b0e5-4255-c38c-f5b61f6a4054"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Mounted at /content/drive\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "import cv2\n",
        "import matplotlib.pyplot as plt\n",
        "import tensorflow as tf"
      ],
      "metadata": {
        "id": "SqtnBHLmMEgP"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "url = \"/content/drive/MyDrive/Colab Notebooks/Google-Consultation/Vesuvius Challenge - Ink Detection/mask.png\"\n",
        "mask = cv2.imread(url, 0)\n",
        "mask = tf.cast(mask, dtype=\"bool\")\n",
        "\n",
        "mask.shape"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "hHgZ4K0qMNTh",
        "outputId": "b78a34d3-6dac-4a30-8f12-de2f78daae85"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "TensorShape([8181, 6330])"
            ]
          },
          "metadata": {},
          "execution_count": 17
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "x = 7000\n",
        "y = 0"
      ],
      "metadata": {
        "id": "4jD_dwCcMaVe"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "ij = tf.stack(tf.meshgrid(\n",
        "    tf.range(x, x+1000, dtype=tf.int64), \n",
        "    tf.range(y, y+4000, dtype=tf.int64),\n",
        "    indexing='ij'), axis=-1\n",
        ")\n",
        "\n",
        "ij.shape"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ZOW9gW84NWis",
        "outputId": "cb9f77ff-e0c4-4f06-9574-6992086fe598"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "TensorShape([1000, 4000, 2])"
            ]
          },
          "metadata": {},
          "execution_count": 35
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "plt.imshow(mask)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 287
        },
        "id": "zpGEVPs1N72p",
        "outputId": "7fcbe2cd-a2e8-46a5-cbc7-2ce7404e4a48"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.image.AxesImage at 0x7f9bedf64ac0>"
            ]
          },
          "metadata": {},
          "execution_count": 36
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAANoAAAD8CAYAAAAR6LrwAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjcuMSwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/bCgiHAAAACXBIWXMAAAsTAAALEwEAmpwYAAAmcUlEQVR4nO2deXhU1f3GP9+ZyR62sMmmRDYBkbAIBBRRiiK0Wm0VbFXcigguFK1oa3+obV1rK1oD4grWShEVrbIouCCy78gaIIBBICBrEhKSmfP74140BJLMTCZz7sycz/PkycyZc+68NzNv7r3nnvMeUUphMBhqFpduAQZDLGCMZjCEAWM0gyEMGKMZDGHAGM1gCAPGaAZDGAi70URkoIhsFpGtIvJQuN/fYNCBhPM+moi4gS3AACAXWAbcoJTaEDYRBoMGwn1E6wFsVUptV0qdAKYCV4dZg8EQdjxhfr9mwHdlnucCPctWEJHhwHAAN+5uydQOnzqDsxFBpSYh+ccpbZAMgBI4p3EeO/c3Im5vQdglHePQAaVUw6rqhdtoVaKUmgRMAqgtaaqn9NesyBAOXCkp+AoKcDeoj9RKheNFbPhrC9q9dJy8x0rx+lxc2iKbvzZewCP7Luf5JsvLtD4bgA4TxtBy6l423d0IleTFc9DDuQ8tqlHdc9X0nf7UC7fRdgMtyjxvbpcZYpTCa3siPsV1f5tD1ruDefK3U2jhOchhXxL9k7zkXp5Pc09qmRaJ5Uz2ExvuymLj7YW0j0/+sWzArFtxfbUKd906eA8fqeG9qZhwG20Z0EZE0rEMNhT4TZg1GByCKzGRz1/MAiBO3NwzPMt+JR7wApQzWdWUNRmAz+Pih4/a0WR0McSK0ZRSpSJyNzAHcAOvK6XWh1ODQT+HbsmkwYK9dJy2gzhx1+h7fTj5JZIknkHbu9bo+1RF2K/RlFIzgZnhfl+DMxCPh0V/e4lNJcV0jE+q8fdLdSWSef8IarO4xt+rMszIEEPNIWL9drkp+FVPJCGB5M/r4RZXWEwG0GfttdR+R6/JwBjNUE3E89NJkbtuHTzntGDvfb1xZXTAN7c5e2e0p/7XtfnihSy6Lynk/dafhVVfrZ/v4uBtmbhr1ybv7t7serQ3++7tXWkb1/nnASAXdkISEtj2bGa1dTiue9/gXDzNm1GauxuV2RnlcVHYJIEO969j2b97UNLvCGM6zKNT4nf0SIgj9/7yvYVu/tpoXdg1J86rz7I2E5j0h6YMrzP/x/LM/SOo8+5yDtxyIY3n7mbr75rhjQdXCcy98Vl+9s0oFlyUxXMH+jC78QQGvX4d3o3ZQesI6xCsQDH30cKMCCgFLjfudufi27qTXWO744uDRitLefKfE7lp/u9Y2v8F3AiproQa78yoKbzKxz3f9yar2WKWFpfQIyGu0vpbSgoY+L8xtLl7ySnlc9X0FUqp7lW9nzGaAQB3u9a4Xi5gx8E0zmu4jyfO/pCsA5f8eM+qWJWQIJV/GWOBLk+MpMm7W/HuywOM0QwB8vyOhafdgzKcmdmFCfyzdXvAf6OZa7QoxJWYyM4HulLasYCEVSk0n3kQ37ebTq2TnAxuN3m/OZ/CxkL7+NV6xEYgA5OL+WeAbYzRIhjfRRnE5+SRc0tLGq8oIX72MgAuXnqIWQ3sURaXwKA3BuBu15pN99anQcuDxLu9ZDbKoVViHiPqfq1xDyKXvJG9aZS10O/6xmgRiislhanvZLGuJJm+iXDllTfA+eex6Z5azGnwyil1b16wnK6JubSNS9GkNvqIzw/sksvcR4tAxOOh6TyhntsyGUDCCwd5eear5PzildPqD611yJgsxIx/9F8B1TdHNAfjqlULVVRMzrhupHXL49iXjSk4p5R+XTfy2tmnnvLNaDMHCGwAriE45hfBrR+OonUAw7qM0RzI7rG9KWzu5aGf/Y+1+S2Y3WyC9UJnvboMFk/3HUy7lAP2/AL/MEZzIN/el/XTkzrf6xNiOCPe/QdQucUBtTHXaA7BXa8ennNbcmxIL91SDFXQdfFxpFvHgNqYI5pD2D2sPf8Z/RzpHjfWxEeDU3mi8Vr2fLCQS97+Azw03a82xmgOoaihCtvUEUP1aeJJJe6o+F3fnDo6hM23TtAtwRAg0+58zu+6xmgOYOe0TrolGAIkz1vA/dcN97t+lUYTkddFJE9Evi1TliYin4lItv27nl0uIvKCHfe9VkS6lmkzzK6fLSLDAtyvqOPkhMmcJzPZdNFbmtUYAuXGISNRy/yfX+fPEe1NYGC5soeAeUqpNsA8+znAlUAb+2c4MAEsYwLjsMJSewDjTpozVmnxTQI//C6TLcPMKWOkkT5jOLJwTUBtqjSaUmo+cLBc8dXAZPvxZOCXZcqnKIvFQF0RaQJcAXymlDqolDoEfMbp5o0Z9s5ozystvmH5Y8ZkkcanhXG0Hbk04HbBXqM1VkrtsR/vBRrbj88U+d2skvLTEJHhIrJcRJaXENhNwUhg7+97s6bHO7plGIJg44lCxl8c3PzIaneGKGvmaMhmjyqlJimluiuluseREKrNOoY1f8iqupLBkdw/6BZK9+wNqm2wRttnnxJi/86zyyuK/DZR4IaIZ+OY4BdcCdZoHwEnew6HAR+WKb/Z7n3sBRyxTzHnAJeLSD27E+RyuyymKPhVT9Ln3K5bhiFIknKCH7FT5cgQEXkH6Ac0EJFcrN7Dp4BpInI7sBO43q4+ExgEbAUKgVsBlFIHReQvWNn7AI8rpcp3sEQ1xYMv5IsXssgtPY6ZzhKZLLrrOa7/W3AZj1UaTSl1QwUvnXZVaF+vjapgO68DrwekLopIXmrlzKfHGZNFKiXKF3RbMzIkTBR3Olu3BEM1SaxGhqUxWhjY8lp3Jr85XrcMQzW5etgZT9b8whithvH268oDmXMCXufL4CzSZ92BZ96KoNsbo9UwM/6dxai631Vd0eBoxvf9T7XaG6PVEK7kZA590oZUV6JuKYYQ0Ctxf7XaG6PVEL0WHWZpl3d1yzCEiEbu6sX1GaOFiGNDeuFKSflx7a1pU/vpFWRwFMZoIcDdthVf/eMlBizZw+qHslB9Mrj6+gW6ZRkchDFaNfA0OQtJSOCJOf8hTtyMSdsOwH+nZvFE47Wa1RlCTfYLPYNua4wWLL0u4L6v59F5cTEZCafOMqjnNssfRSNNvwq+rTFagLhq1UI8Hl6blsXlySU83Xi1bkmGMHHH3z4Iuq0xmp+427eh4Nc9uWbpNhp9nWxuQMcY4/Z3ZNolGUG3N7mOflLctDZfjM8iTtwMNzHdMceSHqmo4ryqK1aAOaL5yZOvTozYhdEN1WffHd2q1d4YzQ/crdPpkWAWSo9lauWWVqu9MZofeLfm6JZg0MxtT8+oVntjND84PiddtwSDZiZu71ut9v4kFbcQkS9EZIOIrBeR++zymEgr3vLKhczvFHy3riE6WJzh36oxFeHPEa0UuF8p1QHoBYwSkQ7ESFpx81nmoG+oPv4kFe9RSq20Hx8DNmKFn0Z9WnH+9b3IHRCyyEpDDBPQfTQRaQl0AZZQg2nFTuGD556r9vQIQ/SQ9+F5pL2QEtRMa7/Pi0QkFXgPGK2UOlr2tVCmFTspEvzpvIu1vr/BWcjsehxsF1x6tl9GE5E4LJO9rZR63y6ukbRiJ0WC7ysOPpnWEH08cf/rFJwd3PHEn15HAV4DNiql/lHmpahOK971aG/+3fJL3TIMDmJgcjFbbp7ArnG9A27rzzVaH+AmYJ2IrLbL/kg0pxW73GwcbhajMJyZhmsCHyXiT1LxAqCiVbGjM63Y59WtwOBgzv/TWrbNCKyNuUlUAQe8BbolGBxKfmngi10Yo1XAhhLTrW84M1ktPg24jTHaGdh7X2/6mjhGQwUEk9VpjFYOd716rBlrOkIMFdP6y1sCbmOMVg5fvrk2M1TMswdb0eo3qwNuZ4xWDlVyQrcEg4N5eU1w02WM0c5Aoc+YzXBmFvf9V1DtjNHKkXd3b9442kq3DINDqeUKbh1rk4JVBk+zpqz6o+kIMZyZjScKmXake1BtzRGtLL7g1yg2VMyWkujoYBo0ezSLewY30NwYrQyle/bqlhBR5HkLaD9xJIe8heSU5FPoO0Hb+TdzxHccgPQ5t5M+6w5GX3kr7d6464zbePZgK7zKx807q5fJEQ48R9z4ioqCaivW0ERnUlvSVE85bThljTJz90rcYv7/+MOMglQmtGmN6t0Z99Fiis9KIW7uCryXduXmCR/x9nnNT6m/8/FMHh0ylcfeuYG4Y+CNh7PHr+bA0M6kvbmU3LE9WX33ixSrEpIknl9tvZLJrWbQbf5dLLs4S/uaBhtPFDK65akj9+eq6SuUUlWeTxqjlWNU9hauSikM63tGKltKCrjnnD5nfM1dPw3vD6dPznDXq4f30KEKt1l4bU/i8r0UNI6j7r8X4+ubgeurVfgu7sKUt19k+rGO/Dx1PYNffZB+V63kyxld+fTOZ0IW0Z7vKyLVlci931/IjfUX8uaBi1m85xzOq5/H4eGN8K7ffEp9Y7QgifuyCR+3nRXW94xE5hfBU4Ovw7sxO2zvKd06Ilt2IS2a4N2wBfF4UKWlSJeOzP7k7YC3d+3WAdzffA5eXLy1vw/ffNSZs5YUk9c9gebPLcfdtDHe3XtQpaXgcp9xVoe/RjO9juXYOTOdben5tIozi1hUxlNXXIs3O3wmA1Ar1lt5GRuOWc9L7Xlh3wauY0ZBKgX9DvJE01+w+fctaD12OS1KFwLQbK6Vy1G6s0zETTWnTpmLkXI0fWYh91x5G4/t76BbiqPp+m54TVYZquQE548fySN5nfxu89jzN4PPS2nubto+tuEn09YQxmhnwLthC18+GPh09Vji1nqLdEs4hWZPL2RFrxQ+Kkhm7L4Mph6rPDJ0/P0/3S/1Hj1aSc3QYIx2BjxnNebdV8frluFoWsWl8t2fe+Np0ZwtWT2saxjN+IqKmNi7D2t6JTClVwa377oIsG4hHPIWMm5/R9rOv5m7d/fk6f5Xh1VblZ0hIpIIzAcSsK7ppiulxolIOjAVqA+sAG5SSp0QkQRgCtAN+AEYopTaYW/rYeB2wAvcq5SqNJxHR2cIwIPb1tE/ycQZ+MPiIi+9Et38MvsKSn7j5od+Z5O26iDeDVt0S0M8Hrb/9UJaPb6G0m7tcC/dgCou/rETJRSErNfRTsFKUUrl27FzC4D7gDHA+0qpqSIyEVijlJogIiOBC5RSI0RkKHCNUmqIHSP+DlYceFNgLtBWKVXhN1qH0Q7cmcmKcRPC+p7Rwvwi6JPgY1dpISNaXgwO7tEOFf4azZ9IcKWUyrefxtk/CrgMOJn8Xz4S/GRU+HSgv23Wq4GpSqlipVQOVkpWD/92Jzy4O7YzJqsGfRPBLS7STY/tafgboOq2o+bysDLztwGHlVInj79l471/jP62Xz+CdXrpVyR4jScV29cS7g5tcSVbIw12PtYbT7OmPPfJG6F/vxjluz9m6pbgKPy6j2af3mWISF3gA+C8mhKklJoETALr1DFU2y35WTdEQcPHc8h5uR1T//Is04524X+7O7Hpgiy+uclH+3i9Q3yiiXrZZoB2WQK6Ya2UOiwiXwCZWKvEeOyjVtl475PR37ki4gHqYHWK+BUJXlN89OZLgBWs4n1yLm5JZWz9bMbWt+4H9Uk0HbCh5PwH1rBjmm4VzsGfSPCG9pEMEUkCBmAt3fQF8Gu7WvlI8JNR4b8GPrdDVT8ChopIgt1j2QZYGqL9qJRDn7Qh1ZX4Y3qRGTRc89SNO65bgqPw54jWBJgsIm4sY05TSn0sIhuAqSLyV2AVVj4/9u+3RGQrcBAYCqCUWi8i04ANWIsbjqqsxzFU5LzTmS1dJldd0RBSDhSnAvlV1osV/IkEX4u1Jlr58u2coddQKVUEXFfBtv4G/C1wmcHjWZ8Cl4TzHQ2TjjRl90XBzduKVqL+HMqbGP33cpzGB9f1rfGxg5FGVBtt/12ZbL7N3BcLN4P+66xxkE4gqo224E9mvGK4af/NTcz6ZTfdMhxHVBut4ydnXD3KUIOk/18R3uztumU4jqg2miSYm6bhZtNdabolOJKoNlrT/5kJ5OHmzks/58iNvXTLcBxRbbTkPaaLOdw8kLaZZx6fqFuG44jqf/nKVdGKwIaaovWc4XQYtxdrzLjhJFF7RJPu5zPtHRPvHW7WX57FkZ6nTcqIeaLWaIMnf609cDMWSXbF88X4LA590gZXolk29SRRa7T/3dFPt4SYJU7cLO3yLn2WHkb1ydAtxxFErdEOdjRHM9080mATv5j0hW4ZjiBqjdZo+gbdEgzAkiPpuiU4gqg0mislhV8t2lx1RUONM/ysL3VLcARRabQeCw9zex2zBJNuDngLeLrvYN0yHEFUGu2rh0zKsBP49Z2jKc0NW1qFo4lKozX4Y45uCQbgV89+qluCY4g6o+1+vyPTW83VLcMAzMrrqFuCY/DbaHa24yoR+dh+ni4iS0Rkq4j8V0Ti7fIE+/lW+/WWZbbxsF2+WUSuCPneAK9mTKmJzRqCYGa7mbolOIZAjmj3YaVfneRp4J9KqdbAIaxMfezfh+zyf9r1sCPBhwIdgYFAlh34E1KeyR0Y6k0aDNXG36Ti5sBg4FX7ueDQSPCNc9uEepOGIDm5aLzB/yPa88CDwMmZlPVxaCR47RwTxuMUfv3bkbolOAZ/AlR/DuQppVaEQQ9KqUlKqe5Kqe5xJATcvt7UsMg0VMHQnMtwfbVKtwzH4M8RrQ9wlYjswFoP7TJgPHYkuF3nTJHg6IgEd6WmhHqThiD4e4uPdEtwFP4s2/SwUqq5UqolVmfG50qp3+LQSHDvoUN4lckK0U1zj1m6qSzVuY82FhhjR3/X59RI8Pp2+RjgIbAiwYGTkeCzqcFI8PcKKl+/2GAIN4GuJvMl8KX92LGR4JclfQ+YU0idtJo2gtYs1i3DMUTdyBCAG24YxcTDZjq9LmYUpNJ6tDFZWaLSaK4Fq3n3e5OWq4Prt/fnlYv76JbhOKIyBSv/+l5808FEnungaP98VHENLIkc4UTlEa3OmgO6JcQs3//3XN0SHElUGs27ZZtuCTGJV/moM7m2bhmOJCqNhjLDsHTQ65FRJH+wRLcMRxKdRgMWF9X4qr2GcqS9aXoaKyJqjZbnraVbQsyhMi/QLcGxRK3RHv3HsKorGUJCTkk+W0oKkIVrdEtxLFFrtF+MmK9bQsxwtieZG//vAd0yHE1UGu3orFY81nC9bhkxg1tc/PGRt3TLcDRRabRFnd/TLSHmeKVfXzxnNdYtw7FE5cgQQ/hRBQV4Dx/RLcOxROURzRBebtzRD1+ByQepDGM0Q7XZ93A6quSEbhmOJuqMVjDbjLULNzsHmgUHqyLqjLbggvd1S4g50h9epFuC4/E313GHiKwTkdUistwuSxORz0Qk2/5dzy4XEXnBTiReKyJdy2xnmF0/W0TMHeUowd3arIFWFYEc0S5VSmUopbrbzx8C5iml2gDz7OcAV2IF77QBhgMTwDImMA7oiRWBMO6kOUPJ9dv7h3qThio4dkEj3RIcT3VOHcsmEpdPKp6iLBZjxdI1Aa4APlNKHVRKHQI+w4oGDynb3mob6k0aqmD3pboVOB9/jaaAT0VkhYgMt8saK6X22I/3AifvVlaUSFzjScXuju1Y/H//CqiNofp48qPuUj/k+PsXukgp1RXrtHCUiPQt+6Kd2xiSSWDVSSqeOOs14kK/boahCr767bO6JTgev4ymlNpt/84DPsC6xtpnnxJi/86zq1eUSFzjScXm/2r4uXbrAPpN+YNuGY7Hn+z9FBGpdfIxcDnwLacmEpdPKr7Z7n3sBRyxTzHnAJeLSD27E+RyuyxkFCoJ5eYMfrD+69a0fMR071eFPweBxsACEVmDFeH9iVJqNvAUMEBEsoGf2c8BZgLbsZZlegUYCaCUOgj8BVhm/zxul4WMX/zbTNUIN/U2mNgIfxDl4HyN2pKmeop/3fWujA58/Mm/cYs5gQwnj+3vwMLO8bplaGOumr6izC2vComab+UrH00yJjM4lqj5Zqa5Yve/qk7GNdygW0JEEDVGM936+tg7uje4zN+/MqLGaOe/ebduCTHLmgezyPvArB1eGVFjtNqdf9AtIWZpO/kurjzbnEJWRtQYLWVCHd0SYpakPGFlryTdMhxN1BjNXWSW09WFq99Bs4JMFUSN0TzzVuiWELMc3m0WtqiKqDGaQR9br54IYoa/VUZUGS23NF+3hJjELS7ctcxaB5URVUZ7+0gX3RJikn7f/hLv0aO6ZTiaqDLaxEX9dEuIOfaU5pNwxU7dMhxPVBnNlVSqW0LMcdH8e8zCj34QVUY778/mpnW4WdD3RVSfDN0yHE9UZe/7UpN1S4gpSpSXW68dgXtjNuYuZuVE1RHN9+0m3RJiisxxd6OWf4uvoEC3FMcTVUbLfqGnbgkxRec71umWEDFEzalj3ofnsf3Cl3XLiCkKvGYOoL/4GwleV0Smi8gmEdkoIplOiwRPnFo3lJsz+MHbLefqlhAx+HvqOB6YrZQ6D+gMbMRhkeClSWYIULjp8g8zB9Bf/ImbqwP0BV4DUEqdUEodxmGR4NeM/jxUmzL4QfuXR9LkuYW6ZUQM/hzR0oH9wBsiskpEXrXzHR0VCb5gyAV+1zVUn5LapkM/EPwxmgfoCkxQSnUBCvjpNBHQHwl+bEgvHp35n1C8vcFPft5vuW4JEYU/RssFcpVSS+zn07GM55hI8LwLoUdCXCg2ZfCTjYfP0i0hoqjSaEqpvcB3ItLOLuoPbMBBkeBtn9wSis0YAmBO+48RT9TcHapx/P1L3QO8LSLxWHHft2KZdJqI3A7sBK63684EBmFFghfadVFKHRSRk5HgEMJIcN9RMw9NB6426Xg3ZuuWERH4ZTSl1GrgTLHHp+V129droyrYzuvA6wHo8wtVciLUmzT4gS/R3LD2l6gagmUIL4fON1kh/hI1Rnskr5NuCTHH8V8d1i0hYogao6V5zAjycFNYkKhbQsQQFUbb+o9ejEnbrltGzLH6kgm6JUQMEW80T/o5bBs6UbeMmCRBzL1Lf4loo7lq1eLxz9/VLSNmMSv4+E9EG63fwr10SzBdzDpxdW6vW0JEELFGy57SlbH1zc1SneSU5MPmHN0yIoLINVr/V3VLiHlGDr4DX1GRbhkRQcQabbFZvEQrHRbeaMKQAiAijVZyeXf6JEak9Kjgxh39aPHrb3XLiCgi7tsqCQl89Pq/dMuIafIeOEe3hIgj4ozmSkok1WVGJOjk6ldNbESgRJzRvIeP6JYQ87zz50G6JUQcEWc0gLUnTE+XTu59cqpuCRFHRBrNp0y0nC5yS/OZfEmmbhkRR0Qa7YbJv9ctIWa5bcgoSvfs1S0j4vAn17GdiKwu83NUREbrSip2ZXTg29+ZXkddpD79vW4JEYk/4TyblVIZSqkMoBtWDsgHaEoqfuWjSbglIg/EUcG2gw10S4hIAv3G9ge2KaV2oimpuK7LJC/pZE2Pd3RLiEgC/dYOBU7+pWssqRjrSEgipy8s6IrMy0pDjOP3t9aOmrsKOG0CWDiTis9/795QvI3BEFYCOTxcCaxUSu2zn2tJKm5+3r6qKxlqlKJf9NAtIeIIxGg38NNpI2hIKs55IhPfK40CaWKoAVK2HdYtIeLw6xrNXj1mAHBnmeKnCHNS8eqbx+MeJoDJqtCJ98XCM0TnGirD36TiAqB+ubIfCHNScYJ4TNe+ZqYcbYAMzAvNBXkMEVHf2vZf36JbQswztVdHE8EeBBFjNImLZ27vLN0yYh5fwXHdEiKSiDFa20WKsz2pumXENIW+EyivV7eMiCRijPZC02VVV3IwHxWcfvO9MvaUWktR7SrN54jvONtK9C9NNejOUeAzRguGiBnPtLS4JGJX9ey05De0+N0+xo5sx8YR1ulvvq+IVFcih7yFfOd1EYePkdk3UD+xgIw6uSy4tRvZo+Np+/civu9Xjzo7S5mfNUnrfiR8Etn/7HQSMUZ7ed+l9Dh7gW4ZQdH0mg14gbMfX0inopEMHLqIlQ905UCnBBqtKCT++8Moj5v47O0cExffxNdFFa2n9U2waWIPlg1+lmM+BZhT50glYoy2cl9zOFu3iuBwt07Hu9UKGm36zELWjU/AU7yCs+zojdKylZUXX9FPp2czBz5PA3cKDRyQvu1u3AjvvryqKxpOIyKu0XIf7s2qCyNz+vzq4mKOdDl1NIsq9j+U8pdLRoRaUvA0CHhWk8EmIo5o6++J3G79P/UfQur2JQG3cyUnc7xfRzZf/EoNqAqcVtNG0Hr9Yt0yIpaIOKJ5lU+3hKApzdkZVLtjgzoxc9JLIVYTPK1HG5NVh4gw2tdFEXHgDSmFDV0ku8xKOdGCo41W0jiFA8MzeWrnlbqlBE3Ok72CarfsEecczQzVx9GHiri8Av78h7e4NGk/kKRbTlA0XhLcaa/TBk9LQkJAnTiGU3HWp1keBZMyezJ49GjdSoLmnfHP6ZZQbfquu8aYrJo422jA1jFteezpyF0LbfDK3wXcRhJOj3DQxSFvIUkDd+iWEfE43mgrb/4n/ZMid3zdou6Tq65UjmNXZTgi9rxYlXD9jaNAmdln1cXR12gAt+0YzLRz5+mWETTJrnhcycn4Cgv9blNnw2EevnQIm0c2ZetvJ1RbQ7EqIUHiyPcVcdhXSrGCmfkd+WjvBexY1hzPcSH+CDx81zsMrXXox3b977ublK8CvwdoOB3HGy37h4Zwrm4V1SPQqSXe9ZsBaPWHnbRyj+C1qyfRL8m/TpW/HjiPe9JW8drhjvxvTycGnrWBTx65lIPDCmiUlUjCgePIiVLUzt24CnM5V/2UADhlcgZ1ly3kkSdvo2jwUZpNNyYLFaL8OC0Qkd8Dd2BFyq3DygFpAkzFijhYAdyklDohIgnAFKxU4x+AIUqpHfZ2HgZuB7zAvUqpSsN5akuaavTUI2wZVv3/6jo59707aXNP8F/anKcy+eo3z9LAnUS+r5gHvx/Ak00/pceMMXTN2Mb6z9riKoWS2opWj66i9ML2eFZsxnf8OK6EhIDWmXY3bIh3//6gtcYac9X0FUqp7lXVq9JoItIMWAB0UEodF5Fp/BTA875SaqqITATWKKUmiMhI4AKl1AgRGQpco5QaIiIdsFK0egBNgblAW6VUhf/ua0uayuw8ilmzI3OcY6t5t/JGnzd4om13VGlp1Q0qwN2xHb5EDwXnpJK0rxjXkm9RXdvD0nXgcps5Yhrx12j+doZ4gCQR8QDJwB7gMmC6/Xr5SPCTPQDTgf4iInb5VKVUsVIqByslq8qAQNkXUFCWY2g1bQStb1rF030HV8tkYJ1KqhXrSX5/CfLNamt7S9dZLxqTRQT+LHKxG/g7sAvLYEewThUPK6VOfoPKxnv/GP1tv34E6/TS70hwEVkuIstLKI7IaRlTj9X7cWxgaW7AGbGGKMSfZZvqYR2N0rFO+VIIYnEKfzlTJPiKYit16cvjLr487mLqMf+na8wuTOCQt5CrsgeSU5LPrtJ8fr7lSvpvuKrKtl7l+/G9A2FKr4yA2xiiG396HX8G5Cil9gOIyPtAH6xVYjz2UatsvPfJ6O9c+1SzDlanSNCR4H++fCh7BpxF0w+2A6CKipnyfhIz282stF2+r4gXMgdQ2rYZsnANozrcjhLBt34zHmDsygxG11/Aiz/0ZmzDJWR8eg9je83iH2t+hmdDCim7FY0+382Gh84i5yorRqBEeYkTaxbml8dd9EvyMS2/DtekHMSHj55P3UfjQwv92S1DDOFPZ0hPrNDTC4HjwJvAcqAv8F6ZzpC1SqksERkFdCrTGXKtUup6EekI/IefOkPmAW2q6gzpKRVE4orw/f2ZrPz9i8SJG6/y4RYXXuVj3YkSWscp+o37PfVfW1TxviUkQIfWsGErJy4+n7i5K3DVqoXv2LHT6ma/1JMOHXdx9PkWHBqWT/HmOrQZn8OuG8+lxasbOXx5OzxFPpI+XFrp39MQXYSs1xFARB4DhmDNul+F1dXfDKt7P80uu1EpVSwiicBbQBfgIDBUKbXd3s6fgNvs7YxWSs2q7H0rNZpN/nU9mfTs89zy2BgO9C6l8Vdu6n+zh+KW9XF/sbLKfQsIETNKwnAKITWaLkTkGLBZt44Q0wA4oFtEDRCN++XPPp2jlGpY1YacPjJksz//LSIJEVkebfsE0blfodwnxw8qNhiiAWM0gyEMON1oeqN5a4Zo3CeIzv0K2T45ujPEYIgWnH5EMxiiAmM0gyEMONZoIjJQRDaLyFYReUi3nsoQkRYi8oWIbBCR9SJyn12eJiKfiUi2/bueXS4i8oK9b2tFpGuZbQ2z62eLyDBd+1RGj1tEVonIx/bzdBFZYmv/r4jE2+UJ9vOt9usty2zjYbt8s4hcoWlXTmqpKyLTRWSTiGwUkcywfE5KKcf9AG5gG9bc6nhgDdZ8OO3aKtDbBOhqP64FbAE6AM8AD9nlDwFP248HAbMAAXoBS+zyNGC7/bue/bie5n0bgzV07mP7+TSs0T4AE4G77McjgYn246HAf+3HHezPLwFrYPo2wK1xfyYDd9iP44G64fictH9JK/hjZAJzyjx/GHhYt64A9H8IDMAa1dLELmuCdQMe4GXghjL1N9uv3wC8XKb8lHoa9qM51pjUy4CP7S/cAcBT/nMC5gCZ9mOPXU/Kf3Zl62nYnzpADnYnYPm/f01+Tk49dfRr7poTsU+ZugBLgMZKqT32S3uBxvbjivbPafv9PPAgcDKwpD41NA8xTKQD+4E37NPhV0UkhTB8Tk41WkQiIqnAe1gDpo+WfU1Z//oi5l6KiPwcyFNKrdCtJYR4gK7ABKVUF6AA61TxR2rqc3Kq0YKeu6YLEYnDMtnbSqn37eJ9ItLEfr0JcHK6eEX756T97gNcJSI7sGZpXAaMx56HaNc50zxEQjUPsQbIBXKVUieTkqZjGa/mPydd5/9VnEt7sC4w0/mpM6Sjbl2V6BWs5K/ny5U/y6kX2c/Yjwdz6kX2Urs8Desaop79kwOkOWD/+vFTZ8i7nNoZMtJ+PIpTO0Om2Y87cmpnyHb0doZ8DbSzHz9qf0Y1/jlp/5JW8gcZhNV7tw34k249VWi9COt0Yy2w2v4ZhHWNMg/Ixkr9SrPrC/CSvW/rgO5ltnUbVnDRVuBW3ftmayprtHOBpba+d4EEuzzRfr7Vfv3cMu3/ZO/rZuBKzfuSgTVxeS0wwzZKjX9OZgiWwRAGnHqNZjBEFcZoBkMYMEYzGMKAMZrBEAaM0QyGMGCMZjCEAWM0gyEM/D/yTCgI99jzAQAAAABJRU5ErkJggg==\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "patch = tf.gather_nd(mask, ij)\n",
        "plt.imshow(patch, cmap=\"gray\")"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 154
        },
        "id": "nFQrUfkNMffj",
        "outputId": "579ac96e-7dcf-46fd-dd25-851a51238d70"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.image.AxesImage at 0x7f9bedfaecd0>"
            ]
          },
          "metadata": {},
          "execution_count": 37
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    }
  ],
  "metadata": {
    "colab": {
      "name": "scratchpad",
      "provenance": [],
      "include_colab_link": true
    },
    "kernelspec": {
      "display_name": "Python 3",
      "name": "python3"
    }
  },
  "nbformat": 4,
  "nbformat_minor": 0
}