Last active
January 9, 2019 09:41
-
-
Save prhbrt/ab544d1834aa41a9e77915d79285a1af to your computer and use it in GitHub Desktop.
Holistically-Nested Edge Detection Paper Example
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
{ | |
"cells": [ | |
{ | |
"cell_type": "code", | |
"execution_count": 1, | |
"metadata": {}, | |
"outputs": [], | |
"source": [ | |
"import numpy\n", | |
"\n", | |
"%matplotlib notebook\n", | |
"from matplotlib import pyplot\n", | |
"\n", | |
"from skimage.io import imread\n", | |
"import caffe\n", | |
"import numpy" | |
] | |
}, | |
{ | |
"cell_type": "code", | |
"execution_count": 2, | |
"metadata": {}, | |
"outputs": [], | |
"source": [ | |
"image = imread('paper-example.jpg')\n", | |
"\n", | |
"# pyplot.figure()\n", | |
"# pyplot.imshow(image)" | |
] | |
}, | |
{ | |
"cell_type": "code", | |
"execution_count": 3, | |
"metadata": {}, | |
"outputs": [], | |
"source": [ | |
"def preprocess_input(X):\n", | |
" X = numpy.array(X, dtype=numpy.float32)\n", | |
" X = X[...,::-1]\n", | |
" X -= numpy.array([[[[104.00698793,116.66876762,122.67891434]]]])\n", | |
" return X.transpose(0, 3, 1, 2) / 128" | |
] | |
}, | |
{ | |
"cell_type": "code", | |
"execution_count": 4, | |
"metadata": {}, | |
"outputs": [], | |
"source": [ | |
"images = image[None]\n", | |
"preprocessed = preprocess_input(images)" | |
] | |
}, | |
{ | |
"cell_type": "code", | |
"execution_count": null, | |
"metadata": {}, | |
"outputs": [ | |
{ | |
"data": { | |
"text/plain": [ | |
"(-0.95842904, 1.1796329)" | |
] | |
}, | |
"execution_count": 5, | |
"metadata": {}, | |
"output_type": "execute_result" | |
} | |
], | |
"source": [ | |
"preprocessed.min(), preprocessed.max()" | |
] | |
}, | |
{ | |
"cell_type": "code", | |
"execution_count": null, | |
"metadata": {}, | |
"outputs": [], | |
"source": [ | |
"net = caffe.Net(\n", | |
" 'deploy.prototxt',\n", | |
" 'hed_pretrained_bsds.caffemodel',\n", | |
" caffe.TEST\n", | |
")" | |
] | |
}, | |
{ | |
"cell_type": "code", | |
"execution_count": null, | |
"metadata": {}, | |
"outputs": [], | |
"source": [ | |
"net.blobs['data'].reshape(*preprocessed.shape)\n", | |
"net.blobs['data'].data[...] = preprocessed\n", | |
"\n", | |
"# run net and take argmax for prediction\n", | |
"net.forward()\n", | |
"out1 = net.blobs['sigmoid-dsn1'].data[0][0]\n", | |
"out2 = net.blobs['sigmoid-dsn2'].data[0][0]\n", | |
"out3 = net.blobs['sigmoid-dsn3'].data[0][0]\n", | |
"out4 = net.blobs['sigmoid-dsn4'].data[0][0]\n", | |
"out5 = net.blobs['sigmoid-dsn5'].data[0][0]\n", | |
"fuse = net.blobs['sigmoid-fuse'].data[0][0]" | |
] | |
}, | |
{ | |
"cell_type": "code", | |
"execution_count": null, | |
"metadata": {}, | |
"outputs": [], | |
"source": [ | |
"# pyplot.rcParams['figure.figsize'] = (9, 5)\n", | |
"# for axis, out, title in zip(\n", | |
"# pyplot.subplots(2,3)[1].ravel(),\n", | |
"# [out1, out2, out3, out4, out5, numpy.zeros(out5.shape)],\n", | |
"# ['out1', 'out2', 'out3', 'out4', 'fusion', ''],\n", | |
"# ):\n", | |
"# if out is not None:\n", | |
"# axis.imshow(out)\n", | |
"# axis.set_xticks([])\n", | |
"# axis.set_yticks([])\n", | |
"# axis.set_xlabel(title)\n", | |
"# pyplot.tight_layout()" | |
] | |
} | |
], | |
"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": "3.6.6" | |
} | |
}, | |
"nbformat": 4, | |
"nbformat_minor": 2 | |
} |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment