Which one generates a better 3D model from video? Luma AI or 3DPresso?

Olli Huttunen
22 Jul 202316:41

TLDRIn this video, the creator compares two AI-assisted 3D modeling applications, Luma AI and 3D Presser, using the latest neural Radiance field (NeRF) technology. The comparison is based on their ability to generate 3D models from video images of various objects captured with a 360-degree camera. The creator details the scanning process, video editing, and the final 3D models produced by each service. Luma AI is praised for its speed and background capture, but 3D Presser delivers more accurate and complete models, making it more suitable for 3D printing or game assets. The video concludes by highlighting the potential of these emerging technologies and encourages viewers to explore these tools themselves.

Takeaways

  • 🤖 There are limited AI-assisted applications for creating 3D models from video images, with Luma AI and 3D Presser being notable examples.
  • 🌐 Luma AI and 3D Presser both utilize the neural Radiance field (NeRF) technology for 3D modeling.
  • 🎥 The user, Olioutron, compares these services by scanning objects with a 360-degree camera and processing the video files through the AI software.
  • 🔍 The 360-degree camera is particularly useful for scanning larger objects, as its wide-angle lens captures more area at once.
  • 📸 During scanning, the user recommends recording three rounds from different angles to ensure comprehensive coverage of the object.
  • 💻 The video material is edited using Insta360 Studio before being uploaded to the AI services for processing.
  • ⏱️ Luma AI processes 3D models faster than 3D Presser, but 3D Presser has introduced a credit-based billing system.
  • 🆓 Luma AI is currently free to use, while 3D Presser offers new users a free credit to start with.
  • 🏺 The statue model produced by Luma AI has a visually impressive Nerf rendering but lacks accuracy in the 3D mesh and texture when exported.
  • 🚀 3D Presser's model of the statue is more accurate and has better textures, making it more suitable for 3D printing or game assets.
  • 🔄 Both AI services have their strengths and limitations, with Luma AI excelling in capturing environments and larger entities, and 3D Presser producing more usable polygon-based mesh models.

Q & A

  • What are the two AI-assisted applications mentioned for creating 3D models from video images?

    -The two AI-assisted applications mentioned are Luma AI, developed by Luma Labs, and 3D Presser, a Korean alternative.

  • What technology do Luma AI and 3D Presser use for creating 3D models?

    -Both Luma AI and 3D Presser use a neural network technology called Radiance Field, also known as NeRF (Neural Radiance Fields).

  • What is the role of a 360 camera in the 3D scanning process described in the script?

    -The 360 camera is used to capture video footage of objects from different angles, which is then processed to create 3D models. It is particularly useful for scanning larger objects or those with challenging surfaces.

  • How does the user ensure they do not appear in the final 3D model?

    -The user recommends standing behind the camera rather than directly at the end of the selfie stick to avoid being visible in the final image.

  • What are the key steps in the 3D scanning process with a 360 camera?

    -The key steps include recording video from three different angles (high, middle, and low), ensuring the camera's horizon lock feature is used, and avoiding self-appearance in the final image.

  • What is the purpose of editing the 360 videos in Insta360 Studio before processing with AI programs?

    -The purpose is to crop the video into a 16x9 format, add keyframes, and rotate the image to keep the object centered during the laps, which improves the quality of the final 3D model.

  • How does Luma AI differ from 3D Presser in terms of processing time and cost?

    -Luma AI is faster to process 3D models than 3D Presser. Luma AI is currently free to use, while 3D Presser operates on a credit-based billing structure, offering new users a certain number of free credits.

  • What are the limitations of Luma AI and 3D Presser in terms of output?

    -Luma AI has limitations on rendering out camera animations, allowing only 30 times per month. 3D Presser, on the other hand, may not always produce usable models, especially for larger objects.

  • Which application produced a more accurate and complete 3D model of the statue?

    -3D Presser produced a more accurate and complete 3D model of the statue, with better textures and a cleaner 3D mesh.

  • What are the potential uses of the 3D models created by Luma AI and 3D Presser?

    -Luma AI is best for saving environments and larger entities, particularly useful in programs like Unreal Engine that understand volume models. 3D Presser is more suitable for creating accurate polygon-based mesh models for use as 3D assets in games or for 3D printing.

  • What is the significance of the development pace of NeRF technology mentioned in the script?

    -The rapid development of NeRF technology suggests that the capabilities and applications of AI-assisted 3D modeling from video images will continue to improve, offering new possibilities for various industries.

Outlines

00:00

📹 Introduction to AI-Assisted 3D Modeling from Video

The video script begins by discussing the limited availability of AI applications that can create 3D models from video images. It mentions Luma AI by Luma Labs and 3D Presser, both utilizing neural Radiance Field (NeRF) technology. The narrator, Oliutron, plans to compare these services using video files of three different objects scanned with a 360 camera. The script explains the process of 3D scanning with a 360 camera, emphasizing the use of the camera's front lens and the importance of positioning oneself correctly during recording. The advantages of using a 360 camera for larger subjects are highlighted, and the script describes the process of recording and scanning objects from different angles.

05:01

💻 Post-Scanning Video Editing and Uploading

The second paragraph details the process of uploading and editing the 360 video footage using Insta360 Studio. The narrator crops the videos into a 16x9 format, adds keyframes, and rotates the images to keep the objects centered. The videos are exported as mp4 files with a high bitrate for better quality. The final phase involves uploading the videos to Luma AI and 3D Presser to see the resulting 3D models. The script compares the speed of processing between the two services and mentions the pricing structures, with Luma AI being free and 3D Presser offering a credit-based system.

10:02

🏆 Comparing the 3D Models and Their Quality

The third paragraph focuses on the comparison of the 3D models produced by Luma AI and 3D Presser. The narrator examines the Statue model and notes that Luma AI's pre-calculated Nerf rendering is impressive but less accurate in 3D mode. The 3D Presser model is found to be more accurate with better textures. The models are further analyzed in Blender, revealing that 3D Presser's model is more compact and less fragmented. The script also discusses the challenges of scanning a lamppost and the trailer, with 3D Presser performing better with the lamppost and Luma AI producing a usable model of the trailer.

15:04

🚀 Final Verdict and Future of AI 3D Modeling

The final paragraph concludes the comparison by stating that 3D Presser is more suitable for creating accurate and complete models, while Luma AI is better for capturing environments and larger entities. The narrator emphasizes that these applications are still in early development and may not always produce working models. The script ends with a call to action for viewers to test these services themselves and reflects on the potential future uses of NeRF technology in 3D modeling.

Mindmap

Keywords

💡AI-assisted application

AI-assisted applications refer to software programs that utilize artificial intelligence to enhance their functionality. In the context of the video, these applications are used to create 3D models from video images, which is a complex task involving image recognition and processing. The video discusses two such applications: Luma AI and 3D Presser.

💡3D modeling

3D modeling is the process of creating a three-dimensional representation of an object or environment using specialized software. It involves capturing the geometry, texture, and lighting of the subject to create a lifelike digital model. In the video, the focus is on comparing the effectiveness of two AI applications in producing accurate and usable 3D models from video images.

💡NeRF (Neural Radiance Fields)

Neural Radiance Fields, or NeRF, is a cutting-edge AI technology used for generating high-fidelity 3D models from a collection of 2D images. It uses deep learning to estimate the volume density and color at each point in space, creating a detailed 3D representation. The video highlights the use of NeRF technology in both Luma AI and 3D Presser.

💡360 camera

A 360 camera is a device capable of capturing images in all directions simultaneously, creating a spherical or panoramic view. In the context of 3D scanning, it can be used to capture an object from multiple angles, which is essential for creating a comprehensive 3D model. The video describes the use of a 360 camera for capturing video footage for 3D modeling purposes.

💡Photogrammetry

Photogrammetry is a technique for creating 3D models from 2D images. It involves analyzing multiple photographs of an object to determine its shape and appearance. The video mentions that transparent surfaces, like glass, can be challenging for photogrammetry modeling, which is relevant to the discussion of the 3D Presser's capabilities.

💡Luma AI

Luma AI is a product developed by Luma Labs, a Silicon Valley startup company. It is an AI-assisted application that uses NeRF technology to create 3D models from video images. The video compares Luma AI's performance with 3D Presser in terms of accuracy and usability of the generated 3D models.

💡3D Presser

3D Presser is a Korean alternative to Luma AI, also using NeRF technology to generate 3D models from video images. The video evaluates its performance against Luma AI, particularly in creating accurate and complete polygon-based mesh models.

💡Polygon model

A polygon model is a 3D model composed of polygons, which are geometric shapes used to approximate the surface of an object. The quality of a polygon model is determined by the number of polygons used, with higher polygon counts resulting in more detailed and accurate models. The video discusses the polygonal structure of the models produced by Luma AI and 3D Presser.

💡Texture mapping

Texture mapping is the process of applying a 2D image or texture onto a 3D model to give it a more realistic appearance. This involves wrapping the texture around the model's surface in a way that matches the model's geometry. The video discusses the texture quality of the models produced by Luma AI and 3D Presser.

💡3D printing

3D printing is a manufacturing process where a 3D model is created by adding material layer by layer, following a digital blueprint. The video discusses the potential use of the 3D models generated by Luma AI and 3D Presser for 3D printing purposes, highlighting the need for accurate and watertight models.

💡Unreal Engine

Unreal Engine is a powerful game engine used for creating video games and other interactive experiences. It supports advanced rendering techniques, including the display of NeRF models. The video mentions that Luma AI's NeRF models are best revealed when used with programs like Unreal Engine that understand volume models.

Highlights

Luma AI and 3D Presser are AI-assisted applications that create 3D models from video images.

Both Luma AI and 3D Presser utilize the neural Radiance field (NeRF) technology.

The user, Olioutron, compares these two web services for 3D scanning from video files.

A 360 camera is used for 3D scanning, capturing 180-degree views of objects.

The 360 camera's white fisheye lens is effective for scanning objects of similar size or larger.

The process involves recording video from multiple angles around the object.

Editing the 360 videos involves cropping and keyframing to keep the object centered.

Luma AI processes 3D models faster than 3D Presser but has limitations on rendering animations.

3D Presser uses a credit-based billing structure, offering free credits to new users.

Luma AI's 3D models have a pre-calculated NeRF rendering animation feature.

The 3D models from Luma AI have lower accuracy and contain holes when translated to polygon models.

3D Presser produces more accurate and complete models with better textures.

3D Presser's models are more suitable for 3D printing and game assets.

Luma AI is better for capturing environments and larger entities with real NeRF models.

The technology is rapidly developing, and the full potential of NeRF models is yet to be discovered.

The user encourages viewers to test these services themselves, even with smartphones.

The video also mentions a DIY camera spinner for 3D scans made from Ikea tools.

Create Stunning AI Art Instantly

No login. No cost. Just incredible AI images !

Create Stunning AI Art InstantlyGenerate Now