DEV Community

[Comment from a deleted post]
Collapse
 
saadmrb profile image
Saad Alem

Done

Collapse
 
kmwill23 profile image
Kevin • Edited

Are you looking to...

  1. Scan for image data within a webpage. Or...

  2. Take a screen capture of part or the whole URL destination?

Also, clarify this part of you could:

'integrate it to "see" the webpage without downloading it'

Taken literally, this would never be possible.

 
saadmrb profile image
Saad Alem

Imagine it as a visual detector, if it finds the information it crawls it, otherwise just pass to another url.

I have an idea ^

 
kmwill23 profile image
Kevin

How do you provide the context for the "information" it is using to determine whether or not to crawl?

 
saadmrb profile image
Saad Alem

I have built a classifier for "images I want to be crawled", the thing is I need to deploy it into the web;

Websites Url -> Detected -> Crawl

 
kmwill23 profile image
Kevin

This "classifier", is it actual image file data, or just contextual information about image data?

Can you provide a sample of this "classifier"?

 
saadmrb profile image
Saad Alem

Image file data, otherwise It would be time consuming to do it all in the same time on the web.

i.e: provide it with images of monkeys, it will search through websites, when it detects the image of the monkey it will be scraped then.

 
kmwill23 profile image
Kevin • Edited

When it is scanning for an image, is it looking for that binary-exact image? Or is it looking for visually similar images?

Are you looking for dynamic web content or static web content?