# Visual Genome Region Description Visualization

In this demo, we will demonstrate how you can get use the Visual Genome Python Driver to get images and also their regions. Next, we will show you how you can visualize the regions to make sure that they work.

#### Getting an image id

First, let's get an image from the dataset.

In [1]:
import matplotlib.pyplot as plt
from matplotlib.patches import Rectangle
from src import api as vg
from PIL import Image as PIL_Image
import requests
from StringIO import StringIO

In [2]:
ids = vg.GetImageIdsInRange(startIndex=0, endIndex=1)
image_id = ids[0]
print "We got an image with id: %d" % image_id

%matplotlib inline

We got an image with id: 1



#### Getting the image data

Next, we will get some data about the image. We specifically want to know the image's url.

In [3]:
image = vg.GetImageData(id=image_id)
print "The url of the image is: %s" % image.url

The url of the image is: https://cs.stanford.edu/people/rak248/VG_100K_2/1.jpg



#### Getting the region descriptions

Now, let's get all the region descriptions for this image.

In [4]:
regions = vg.GetRegionDescriptionsOfImage(id=image_id)
print "The first region descriptions is: %s" % regions[0].phrase
print "It is located in a bounding box specified by x:%d, y:%d, width:%d, height:%d" % (regions[0].x, regions[0].y, regions[0].width, regions[0].height)

The first region descriptions is: the clock is green in colour
It is located in a bounding box specified by x:421, y:57, width:82, height:139



#### Visualizing some regions

Now, we will visualize some of the regions. The x,y coordinates of a region refer to the top left corner of the region. Since there are many regions, we will only visualize the first 8.

In [5]:
fig = plt.gcf()
fig.set_size_inches(18.5, 10.5)
def visualize_regions(image, regions):
response = requests.get(image.url)
img = PIL_Image.open(StringIO(response.content))
plt.imshow(img)
ax = plt.gca()
for region in regions: