Create 3D model from photos

With rooom, you can automatically create a 3D model from photos. Depending on the number of images, either an AI-based solution or classic photogrammetry is used.

When is which pipeline used?

When you upload images, our system automatically decides which method to use:

1. For up to 10 images, the AI-based pipeline is used. It is fast, efficient, and ideal for simple objects or when only a few images are available.
2. For more than 10 images, the system uses classic photogrammetry, which calculates precise 3D geometries from multiple perspectives.

Tips for AI-Based Model Generation

When using up to 10 images, the AI-based generation pipeline is applied. To ensure optimal results, please follow these guidelines:

• Use images of the exact same product in an unchanged state – no different versions, orientations (e.g. open vs. closed), or deformed conditions.
  
• The product should appear identical in all images: shape, color, surface texture, and materials must be consistent.
• Avoid product variants or different production series, even if they look similar on the outside.
• Capture images from various but meaningful angles (e.g. front, diagonal, side, top). This helps the AI better understand the overall shape.
• Avoid extreme perspectives (e.g. very close-up or from below) that don’t provide useful information for the model.
• Avoid shadows or reflections that could distort the shape.
• Use even, natural lighting to prevent color distortion and harsh contrasts.

The goal is to provide the AI with a clear and consistent representation of the product. The more uniform your image data, the more realistic and high-quality the resulting 3D model will be.

Tips for model generation with the photogrammetry pipeline

For classic photogrammetry, ideally capture between 20 and 60 images. Keep the following in mind:

• The object should be well-lit and visible from all angles.
• Use a high-quality camera – avoid smartphone cameras.
• For larger objects: use a camera with a 50 mm lens.
• For small to medium-sized objects: use a 70 mm or 100 mm lens.

• Also take close-up shots (e.g., logos, textures).

   

Upload photos and generate 3D model

1. Go to the product overview and click on the tile with the +.
   
2. Drag and drop your images (.jpg or .jpeg) into the upload window or select them manually. Alternatively, you can upload a ZIP archive containing all images.
   
3. After the upload, the system automatically generates a 3D model from the images.

If needed, you can request professional post-processing – click here for more information.

Tips for uploading

• Maximum file size per upload: 100 MB.
• For organized uploads, use a ZIP archive (recommended for many files).
• How to create a ZIP archive:
  • Windows: Select files > Right-click > "Send to" > "Compressed (zipped) folder".
  • Mac: Select files > Hold CTRL > Right-click > "Compress".