2D images quickly transformed into 3D thanks to revolutionary AI

Men can see a two-dimensional image and imagine what three-dimensional really is, but machines cannot. However, a new device invented by scientists at MIT can transform three-dimensional images from 2D images. The tool is approximately 15,000 times faster than older devices.

MIT News

Scientists have been able to use neural systems to determine life-size images from 2D images. However, this machine learning method is too slow to be applicable in multiple labs.

A sensitive device communicating with its surroundings must be able to determine one of the 3D images from 2D images. Rays of light developed by scientists can reproduce a ray of light after a single observation of an image.

How does it work ?

The technique reproduces the images by means of a set of light rays at 360 degrees, passing through all the points and following all the axes. The set of rays is encoded in a neural channel, which makes it possible to achieve the desired result much more quickly.

When a computer wanted to do this, obtaining a 3D object from an image involves mapping many rays of light sent by the camera. Currently, it is possible to do this, but this process involves a lot of calculations and therefore a long wait time.

A light field network (LFN) can transform the light beam of a 3D image. Then, associate with each of the camera rays in the light beam the color seen by this ray. However, to reconstruct all the rays, the neural network must first identify the materials of the light rays.

Therefore, the scientists put their device to the test using various two-dimensional images. As soon as the example learned the shape of a light field, it was able to transform a 3D image from a single image.

Faster renderings?

The scientists experimented with their device by reconstructing 360-degree beams of light from many simple images. They found that LFNs can render scenes at a speed of around 500 frames per second. Also, the 3D objects rendered by the LFNs are much clearer than those rendered by the old systems.

An LFN also requires less storage memory, around 1.6 megabytes, or 146 megabytes for a much older device. In the future, scientists would like to have a more efficient device to be able to use it in slightly more complex situations than today.


PHP Script, Elementor Pro Weadown, WordPress Theme, Fs Poster Plugin Nulled, Newspaper – News & WooCommerce WordPress Theme, Wordfence Premium Nulled, Dokan Pro Nulled, Plugins, Elementor Pro Weadown, Astra Pro Nulled, Premium Addons for Elementor, Yoast Nulled, Flatsome Nulled, Woocommerce Custom Product Ad, Wpml Nulled,Woodmart Theme Nulled, PW WooCommerce Gift Cards Pro Nulled, Avada 7.4 Nulled, Newspaper 11.2, Jannah Nulled, Jnews 8.1.0 Nulled, WP Reset Pro, Woodmart Theme Nulled, Business Consulting Nulled, Rank Math Seo Pro Weadown, Slider Revolution Nulled, Consulting 6.1.4 Nulled, WeaPlay, Nulledfire

Back to top button