Street Scene 3D Reconstruction

Upload an image or select an example to see:

  • Semantic Segmentation - Identify roads, buildings, vehicles, people, and other scene elements
  • Depth Estimation - Generate metric depth maps showing distance to objects
  • 3D Point Cloud - Interactive 3D reconstruction with semantic colors)

CPU Mode

⚠️ Upload disabled: DepthAnythingV2 requires CUDA. Using precomputed examples only.

Examples

Click on an image to load the example:

Results

Semantic Segmentation

Depth Estimation

3D Point Cloud