Created
November 22, 2018 08:01
-
-
Save CasiaFan/038e25ee10bcc00e4ed0791f1cd3fa3b to your computer and use it in GitHub Desktop.
An example of using tensorflow hub for image classification
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
from urllib.request import urlopen | |
from io import BytesIO | |
from PIL import Image, ImageOps | |
import cv2 | |
def download_and_resize_image(url, filename, new_width=256, new_height=256): | |
response = urlopen(url) | |
image_data = response.read() | |
image_data = BytesIO(image_data) | |
pil_image = Image.open(image_data) | |
pil_image = ImageOps.fit(pil_image, (new_width, new_height), Image.ANTIALIAS) | |
pil_image_rgb = pil_image.convert('RGB') | |
pil_image_rgb.save(filename, format='JPEG', quality=90) | |
# instantiating module | |
tf.reset_default_graph() | |
module = hub.Module("https://tfhub.dev/google/imagenet/inception_v3/classification/1") | |
height, width = hub.get_expected_image_size(module) | |
img_url = "https://cdn.pixabay.com/photo/2017/02/20/18/03/cat-2083492_960_720.jpg" | |
img_jpg = "cat.jpg" | |
| |
# input and output tensor | |
input_tensor = tf.placeholder(tf.float32, shape=[1, width, height, 3]) | |
logits = module(input_tensor) | |
labels = tf.argmax(logits, axis=1) | |
# download_and_resize_image(img_url, img_jpg, width, height) | |
image = cv2.imread(img_jpg) | |
# input range from 0-1 | |
image_rgb = (cv2.cvtColor(image, cv2.COLOR_BGR2RGB)/256) | |
inputs = np.expand_dims(image_rgb, 0).astype(np.float32) | |
# set up session | |
initializer = tf.global_variables_initializer() | |
sess = tf.Session() | |
sess.run(initializer) | |
# predict | |
res = sess.run(labels, feed_dict={input_tensor: inputs}) | |
print(res) | |
# 283: 'Persian cat', |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment