Read time: 4.1 minutes
3D Models market size was valued at USD 1255.8 million in 2022 and is expected to expand at a CAGR of 19.04% during the forecast period, reaching USD 3573.08 million by 2028.
What it has to do with WebAR? Well everything. 3D model is the heart of AR experience. Since it's a new technology, many people think it's easy to create them like just taking a photo using a phone. Let's debunk some myths today.
——————————————————————————————
Neural Radiance Field or NeRF
NeRFs are getting massive popular now a days. Capture anything from your mobile and create a 3D model out of it easily. The introduction of Apple LIDAR (Light Detection and Ranging) scanning fueled the growth of such technologies. In general, NeRF is a fully-connected neural network that can generate novel views of complex 3D scenes, based on a partial set of 2D images.
Example: Luma.ai is a great tool in 2023 to capture 3D simply via capture scenes from phone. There are more such tools are coming up recently in the market.
Limitation: Although NeRF based 3D models are great but if you need to use 3D models for professional WebAR usage (such as retail objects), you need to brush up the 3D model still in 3D editor like Blender to shed the unnecessary artifacts.
Generative AI
Create 3D models by AI software simply by providing text inputs – isn't it amazing? Very few software can actually create a professional level 3D output without deviating much from the actual description yet. Generative AI is developing rapidly and we do expect to see some wonder.
Example: Masterpiece Studio is focusing on to add value to gaming industry by creating 3D models from text input.
Limitation: Most often AI still has hard time to understand and produce the desired resutls. For ecommerce and retail usage, it may not be the ideal platform to use.
When Generative AI and NeRF combine
University of California, Berkeley researchers have launched a combined platform this week combining these two technological titans. It can capture real world objects, convert into 3D and then based user input can alter the 3D – a wonderful approach. See the amazing video here.
ChatGPT for Blender or BlenderGPT
Well in a ChatGPT4 crazed world, how can we keep it aside while talking about 3D modelling. ChatGPT can enable designers to put text input to Blender (3D modelling software) and create 3D models without any manual designing phase. Another marvel in the progress. Read here.
But overall, one thing is clear – all sorts of 3D model generation processes haven't been matured so much that can obselete 100% of human touch. If you are going to use them for advertising, marketing, ecommerce or retail industries – you still need 3D desginers to finish the up the quality, specs, textures, lighting properly for a professional presentation.
If you want to get the complete list of all 3D capturing apps to fast track reality capture method, you can find them inside my book.