A disparity image is basically a depthmap that’s calculated by comparing two stereoscopic images (one from the left and one from the right).
- If there’s a big difference in movement between the two (disparity), it means that point is close to the viewer (light).
- If there’s little movement, it means the point is far away (dark).
Here’s an example with both inputs and the resulting depthmap below:
I’ve seen lots of videos and images that show this kind of calculation but I haven’t seen any web-based examples that run on the browser. There are some cool implementations that are proprietary. Does anybody know of a WebGL or other open-source demo of this that I could look at so I could port to Three.js? I understand it’s a big request, but I just thought I’d ask if anybody knows of an example of this that I could use as a starting-off point.