Home Artificial Intelligence “3D modeling is the key media of the long run”

“3D modeling is the key media of the long run”

1
“3D modeling is the key media of the long run”

One of the talked about words today is ‘generative artificial intelligence (AI)’. ‘ChatGPT’ can be a sort of generative AI. By the best way, there’s a domestic company that has already began developing technology at a time when this word was unfamiliar.

It’s Recon Labs, generally known as a 3D modeling powerhouse. This company began official service early last yr ahead of ‘Dali’ and ‘Midjourney’ and has established itself as a number one company on this field in Korea in only one yr. In other words, it’s a pacesetter within the commercialization of domestically generated AI.

Recon Labs CEO Ban Seong-hoon, whom we met on the CES held in Las Vegas in January, was busy preparing for overseas expansion. Meeting him again, he heard about overseas expansion strategies and business directions using generative AI. [편집자주]

Ban Seong-hoon, CEO of Recon Labs (Photo=Reporter Lee Joo-young)

“I’ll participate as a 3D technology representative on the NVIDIA Developer Conference ‘GTC’, which shall be held from the twenty fourth. I shall be in command of this field session.”

CEO Ban Seong-hun cited technology that’s recognized overseas as the best strength of Recon Labs.

In actual fact, this company was recently chosen for the ‘Inception Program’, a partnership that Nvidia has with the perfect revolutionary startups in major fields resembling artificial intelligence (AI) and massive date. It also entered right into a service provision partnership with Sandbox, a worldwide metabus company.

“We plan to release a recent product, ‘3D Presso’, based on ‘Metaricon’, the technology that made Recon Labs what it’s today, inside this month. It incorporates the know-how accrued over the past 3 years and our will to advance overseas.”

CEO Ban viewed the potential within the overseas market very positively. It is alleged that the likelihood was confirmed at CES 2023, which was participated firstly of the yr. It’s in the identical context that each one menus of this recent product are made in English.

The service area has been expanded to areas for 3D creators resembling games, video, and metaverse.

As well as, it has added a function that may convert the result into various styles beyond scanning objects to create 3D models. It also provides a community function that enables users to share their findings and insights.

CEO Ban explained, “We plan to take part in the Game Developers Conference (GDC) to be held within the US in May and open our business in earnest.” .

Recon Labs demonstrated Metarecon technology on the last CES site, drawing attention from firms in various fields resembling e-commerce, augmented reality (AR) and virtual reality (VR), autonomous driving, digital twin, animation, and games.

It was due to the differentiated performance that Metarecon, which completes real-life 3D assets without using separate equipment, is clearly distinguished from other 3D construction services.

“The demand for 3D modeling will proceed to grow. The performance of AR/VR headsets, including the metaverse, is steadily evolving, and the computing power to implement 3D models is increasing at an astounding rate. It is definite that 3D modeling will emerge as a significant media eventually. do.”

Ban predicted the long run of 3D modeling as follows. In actual fact, he got into the 3D modeling business because, paradoxically, he was well aware that 3D work was laborious and time-consuming.

He has been lively as a VR media producer since his days as a graduate student at Korea Advanced Institute of Science and Technology (KAIST) Culture Technology (CT). He also lectured at universities. Then, he founded Recon Labs with the will to create a service that may easily produce 3D content. That is the background of his founding.

The incontrovertible fact that the core technology of Recon Labs is the technology of implementing objects through 3D modeling shouldn’t be irrelevant to the background of its establishment.

Ban was well aware of the constraints that easy 3D scanning alone couldn’t perfectly reproduce an asset. So, we added NeRF (Neural Radiance Field), a generative AI technology that was just emerging on the time.

Through this, it was possible to complement 3D assets and add technology and AR functions to create a screen that will be rotated 360 degrees. The sector that may best apply this function to the actual industry is e-commerce, which Recon Labs has been specializing in thus far.

The business model developed in this fashion was ‘Flicka’, which was launched early last yr. It’s a service that creates a 3D asset by uploading an object photographed with a cell phone to the online, shows the product in three dimensions, or incorporates additional information. We were in a position to elicit great responses from e-commerce firms that desired to create digital showrooms or catalogs.

The 3D and AR virtual exhibition halls displayed on the ‘Body Friend’ website were also built with Flicka. The variety of service customers that began in this fashion has increased to about 80 firms inside a yr. We’re currently discussing with KT Commerce to use 3D AR.

Flicka’s strength is, above all, the standard of 3D assets which can be indistinguishable from live motion. CEO Ban said confidently, “Other firms have similar technologies, but we’re the perfect by way of quality.”

Based on its technology, the corporate attracted a complete investment of 5.5 billion won from Naver, Kakao, and Lotte in May last yr. The variety of team members, which had been around 10, has increased to about 40 people without realizing it.

In the long run, it plans to release generative AIs resembling ‘Dali’ and ‘Midjourney’.

To this end, we’re developing a ‘Text to 3D’ generation AI that creates a 3D model when the user inputs the specified content in words. As well as, the ‘Sketch to 3D’ model, which implements a hand-drawn picture right into a 3D model, is being promoted as a government research project.

Regarding this, he said, “The ‘text-to-text’ generation AI technology has already been proven with ChatGPT,” and added, “The goal is to implement ‘anising to 3D’, which creates anything resembling sketches, voices, texts, and photos in 3D.” said.

At the identical time, he said, “The worth of 3D model creation technology will grow in a distinct dimension in the long run, and ‘replica’ shall be at its core.”

Reporter Juyoung Lee juyoung09@aitimes.com

1 COMMENT

LEAVE A REPLY

Please enter your comment!
Please enter your name here