Google Image scraper

akashgshastri · September 28, 2020, 9:30pm

Ever found it hard to make datasets for deep learning image/ computer vision projects?
Well worry no more, I wrote this convenient script to automatically scrape images from google for each class of image you might want. Check it out on my GitHub and let me know how you like it.
Link:

github.com

akash-shastri/Chrome_image_scraper/blob/master/Chrome_Image_scraper.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Web Scraper to make Image datasets\n",
    "\n",
    "This is a small script to scrape images from google searching for a particular term, and have images of each term in a separate folder. This is extremely useful if you want to do deep learning with images, and need images from google to make a dataset.\n",
    "\n",
    "Idea and starter Credits : https://towardsdatascience.com/image-scraping-with-python-a96feda8af2d\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Things to do before this script works\n",
    "1. Install selenium (with conda you can use \"pip install selenium\") \n",

This file has been truncated. show original

amitness · October 5, 2020, 4:09am

Very interesting. I used to rely on this library:

jimmiemunyi · October 5, 2020, 6:03am

https://joedockrill.github.io/jmd_imagescraper/ by @joedockrill works well too out of the box too.
Scrapes images from DuckDuckGo

joedockrill · October 5, 2020, 1:10pm

Selenium scrapers won’t work on Colab/paperspace etc unfortunately (which a lot of us use).

akashgshastri · October 9, 2020, 8:11pm

You could run script locally and upload dataset, maybe to kaggle or colab.

iskode · March 8, 2021, 10:34pm

I’ve tried @akashgshastri solution and it works very well on colab.

drscotthawley · March 27, 2021, 1:22am

Since @akashgshastri’s solution didn’t work for me on Colab, and @iskode was a little brief on the details,…

Here’s a link to my working Colab notebook with my modified version of Akash’s notebook, that uses selenium (@joedockrill) via a very nice package called “kora”!

Mozelle · April 14, 2021, 10:18am

I am facing the same problem. Please keep sharing this type of post ahead. Looking for the same issue.
Thanks in advance,

official site

Yep · November 12, 2021, 12:23pm

I found your solution on Google search. Thanks for sharing. I also tried a web service for it and it works quite good.

https://outscraper.com/google-images-scraper/