CSM's Cube Revolutionizes 3D Asset Creation with AI-Assisted Workflows
CSM Company's innovative Cube platform revolutionizes 3D asset creation through sophisticated AI-assisted workflows that bridge traditional 3D artistry and digital asset generation. By integrating advanced AI tools with industry-standard 3D processes, CSM enables artists of all skill levels to produce game and simulation-ready assets with unprecedented efficiency. This comprehensive guide explores the company's unique approach to 3D content creation, from initial asset preparation to final model export, demonstrating how CSM is transforming the 3D production landscape through its intuitive platform and cutting-edge technology.
CSM Company employs a collaborative approach to 3D asset creation, integrating their Cube platform into workflows alongside traditional 3D artists and developers. Experiments have demonstrated significant improvements in production efficiency through this integrated approach.
The company's workflow begins with the creation of game and simulation engine-ready assets featuring high-fidelity geometry, topology, parts, textures, and UV maps. This process typically involves breaking down input geometry into constituent parts or generating parts separately before assembly. Each component can then undergo individual texturing before being processed through their AI-assisted UV remapper. Final meshes are exported to third-party editors for additional refinement.
To leverage Cube's capabilities, artists follow a structured seven-step process:
Utilizing one or more CSM tools
Exporting outputs to third-party tools (Blender, ZBrush, Substance, etc.)
Breaking down input geometry into parts
Assembling these parts
Retexturing each whole or part
Processing through AI-assisted UV remapper
Exporting to third-party editors for final refinement
The platform supports diverse asset types, including humanoid characters, mechanical parts, and environmental elements. Example successful inputs encompass white polar bears, baby astronauts, and wooden buckets, demonstrating its versatility across multiple industries. The company's approach enables both beginners and experienced artists to create high-quality 3D assets with minimal existing 3D software experience.
The production process begins by breaking down input geometry into constituent parts or generating parts separately before assembly, a method that particularly benefits complex object creation. Artists work through a structured seven-step process facilitated by Cube's AI capabilities, enabling both new and experienced 3D practitioners to achieve professional results with minimal existing software experience.
The workflow's foundation rests on uploading high-resolution imagery to the Image to 3D tool, where users can leverage Manual Segmentation for refined object extraction and control over segmentation precision. Multiple refinement options enable optimization between processing speed and detail accuracy, with settings tailored for both slow and fast refinement processes. This flexibility allows artists to balance between expedited workflow and precise object extraction requirements.
Cube's AI platform specializes in creating custom 3D assets for games and simulations, particularly within Unreal Engine's Fortnite Editor (UEFN). The system employs two primary methods for generating assets from 2D images: the Image to 3D tool and access to online marketplaces featuring pre-made assets.
The Image to 3D tool enables users to create 3D models by uploading images to 3d.csm.ai. To optimize the process, artists can enable Manual Segmentation for precise object extraction and control over segmentation detail. The tool offers multiple refinement options, allowing users to balance processing speed against detail accuracy through settings tailored for both slow and fast refine processes.
For example, when running the fast refine option, artists can select specific topology types and configure polycount settings to achieve the desired balance between detail and optimization. Each uploaded image should clearly display the object of interest; for optimal results, users should ensure the subject is well-isolated and free from complex backgrounds or multiple overlapping elements.
CSM's workflow begins with artists uploading high-resolution imagery to the Image to 3D tool or purchasing pre-made assets from their online marketplaces. The company's workflow consists of seven key steps:
Selecting one or more CSM tools
Exporting outputs to third-party tools (including Blender, ZBrush, Substance, and other 3D editors)
Breaking down input geometry into constituent parts or generating parts separately
Assembling these parts into the final 3D model
Retexturing each whole or part element
Running through an AI-assisted UV remapper to optimize texture placement
Exporting the final mesh to third-party editors for additional refinement
This structured approach supports both new and experienced 3D practitioners in creating professional-grade assets with minimal existing software experience. The platform's capabilities span multiple industries, successfully converting diverse input items such as white polar bears, baby astronauts, and wooden buckets into usable 3D assets.
To generate the best possible results, artists should prepare their input images carefully. The system performs most effectively with subjects in T-pose, clearly visible in the foreground with few surrounding elements. For complex subjects like medieval architecture or detailed character rigs, the tool requires specific preparation techniques including manual segmentation to preserve fine details.
The platform also includes advanced features like texture retouching through the Retexture tool, allowing artists to paint colors, AI textures, or original input images directly onto their 3D models. Once created, models can be downloaded in FBX format and imported into Unreal Engine or Roblox Studio for further customization and integration into game projects.
The model creation process begins with selecting appropriate input images, which should display the object clearly in the foreground with minimal background complexity. For optimal results, artists should prepare their images with the subject in T-pose, as this position provides the best balance of visibility and structural consistency.
CSM's tools support a wide range of input types, including human characters, animals, mechanical parts, and environmental elements. Successful examples include white polar bears, baby astronauts, and wooden buckets, demonstrating the platform's versatility across multiple industries. While the system works best with well-isolated subjects in clear focus, it may struggle with stylized illustrations, complex backgrounds, or multiple overlapping elements.
Artists can enhance their results through the platform's segmentation tools. For precise object extraction, users can enable Manual Segmentation, which allows for more controlled segmentation and tighter color control. This feature is particularly valuable for separating complex objects or maintaining fine details during the extraction process.
Once objects are extracted, the platform offers multiple refinement options to balance processing speed and detail accuracy. Users can choose between slow and fast refinement processes, each with its own set of configuration options. The topology type selection allows artists to optimize between mesh complexity and processing efficiency, while polycount settings enable control over the final mesh's detail level.
For post-processing, CSM's tools include a Retexture function that allows artists to directly paint colors, AI textures, or original input images onto their 3D models. This feature provides additional creative flexibility while maintaining the efficiency of the platform's core workflows.
The final output is available in FBX format, ready for integration into both Unreal Engine and Roblox projects. Artists can upload their models to the platform, customize names and positions using intuitive shift, scale, and rotate controls in Roblox Studio, and incorporate their assets into game projects with minimal additional setup.
CSM offers both custom asset creation through their Image to 3D tool and pre-made asset acquisition via their marketplace system. The Image to 3D tool enables users to create custom 3D models by uploading images to 3d.csm.ai. For improved control and segmentation accuracy, users can enable Manual Segmentation and configure multiple refinement variations.
The platform provides options for both slow and fast refinement processes, allowing artists to balance processing speed against detail accuracy. When using fast refine settings, users can customize topology types and polycount settings to optimize between mesh complexity and processing efficiency.
In addition to custom asset creation, CSM maintains extensive marketplaces featuring pre-made assets. These include offerings in the Unreal Engine Marketplace and other online stores, providing users with a wide selection of ready-to-use 3D models for various applications.
The company supports diverse input types, successfully converting images of humans, animals, mechanical parts, and environmental elements into 3D assets. Examples include white polar bears, baby astronauts, and wooden buckets, demonstrating the platform's versatility across multiple industries.
For the best results, artists should upload high-resolution images displaying the object clearly in the foreground with minimal background complexity. The system performs most effectively with subjects in T-pose and well-isolated from other elements. The platform includes advanced features like texture retouching through the Retexture tool, allowing artists to directly paint colors, AI textures, or original input images onto their 3D models. All created assets can be exported in FBX format for integration into Unreal Engine or Roblox Studio projects.