about computer vision | Coursera Community
Coursera Header

about computer vision

  • 19 October 2019
  • 1 reply
  • 101 views

Badge +1

Hello everyone,

i am a beginner and i don’t know if i have enough knowledge to even ask this question.

question is that i i have thousand object in my room do i have to train a thousand classifier with million of labelled images. and this seems hell of a task to me. i know image processing algorithmscally very complicated task . if anyone can tell if there any other way possible to do real image classification rather than having to train with thousands of images!

thanks 


1 reply

Userlevel 1
Badge +1

My knowledge in the latest advancements in the Computer Vision field is quite shallow, but it feels to me that you can only classify positively something you have trained your model for beforehand.

However, you don't need to train 1000 separated classifiers. You can train one model that will output multiple classes, rather than just one.

This article will tell you more about it: https://www.depends-on-the-definition.com/guide-to-multi-label-classification-with-neural-networks/

Reply