Dataset processing

r
feature_engineering
json

#1

I have a dataset in Json format check screen shot:

so if you can check the ImUrl column has images links, and I want to process the dataset so as I can have the images in my local computer, to have a dataset that look like the following:


#2

Hi @erigits, Correct me if I am wrong, do you want to convert the string “http://ecx.images …” to “/data/imagenet” ? Or do you want to open each individual image, and find its width, height?


#3

@jalFaizy the second image is a sample dataset screenshot, I would like my dataset to have, so i want to have the image imUrl column plus yes to open each individual image, and find its width, height?


#4

@erigits, you have to open it through code. For example, in python you could use scipy to open image an extract its height, width.

from scipy.misc import imread

image = imread(image_path) 

print type(image)
<type 'numpy.ndarray'>

print image.shape
(32, 32)