search

hiukim / mind-ar-js

Web Augmented Reality. Image Tracking, Face Tracking. Tensorflow.js
MIT License
2.24k stars 417 forks source link

Making "target" images #177

Closed prp-e closed 2 years ago

prp-e commented 2 years ago

Greetings. I have a question, how target images are built? what do they have inside them which matters? Thanks.

shabarito commented 2 years ago

Hello, Have you checked this link?

prp-e commented 2 years ago

Hello. Yes, I did. But I still couldn't find out how to create target images on my own.

prp-e commented 2 years ago

I explored the codes of this repository and I found that the compiler uses OpenCV. If I know what functions are used, I may be able to make a target image compiler in python as well.

@hiukim

hiukim commented 2 years ago

If memory served me well, I think the compiler didn't rely on OpenCV. It requires webgl though. If you want to make one in python, you will need to rewrite the whole logic (which is currently a gpu code), so that is actually not straight forward. but feel free to go ahead.

I'm also closing this issue for now. Feel free to re-open if there is some new insight.