MATRIX Inc. (hereinafter “MATRIX”), a subsidiary of MetaReal Corp., both headquartered in Tokyo’s Chiyoda-ku, announced on March 27 that MATRIX will start providing an AI-powered contract development solution that can easily generate photorealistic digital twins (*1) from videos taken with smartphones through Gaussian Splatting (*2), a next-generation 3D spatial construction technology. It is intended for use in industries such as construction, engineering, real estate, and manufacturing.

Existing methods for constructing digital twins use a combination of 3D laser scanning and planar captured images, but the former has problems in terms of time and cost, and the latter has limitations in terms of perspective and field of view. Using the latest Gaussian Splatting technology to automatically generate realistic digital twins from video footage taken with just a smartphone — without any special equipment or effort — overcomes these hurdles.

MATRIX has developed and operates Metaverse space-sharing platform “DOKODEMO Door,” which features real 360-degree videos. Through the operation of this platform, the company has received various requests and inquiries about Metaverse development from corporate clients, and this AI technology envisions the following use cases as examples:

Examples:

https://cdn.kyodonewsprwire.jp/prwfile/release/M107827/202403218298/_prw_PI1fl_D971Bco5.png

https://cdn.kyodonewsprwire.jp/prwfile/release/M107827/202403218298/_prw_PI2fl_SuDUi10P.png

https://cdn.kyodonewsprwire.jp/prwfile/release/M107827/202403218298/_prw_PI3fl_at9qhA2W.png

https://cdn.kyodonewsprwire.jp/prwfile/release/M107827/202403218298/_prw_PI4fl_284piAVo.png

Inquiries about development contract development: https://www.matrix.inc/en 

Notes:
(*1) A technology that reproduces environments and objects that exist in the real world in a VR environment

(*2) Gaussian Splatting is an AI-based 3D spatial reconstruction method announced in 2023 that can more easily and quickly generate 3D images than existing technologies such as photogrammetry and NeRF.

Leave a Reply

Your email address will not be published. Required fields are marked *