Available now

Marigold Depth Estimation: depth map

Marigold Depth Estimation turns a single photo into a depth map that shows how far each part of the scene is from the camera.

Model ID

Input

Fields marked * are required. Review the price before generating.

Uploading a file needs an account. Create account

JPEG, PNG, WebP · up to 25 MB

Additional settings

Number of denoising steps. Defaults to 10. The higher the number, the more accurate the result, but the slower the inference.

Number of predictions to average over. Defaults to 10. The higher the number, the more accurate the result, but the slower the inference.

Maximum processing resolution. Defaults 0 which means it uses the size of the input image.

Price$0.28 / run

You sign up first. Your settings are kept.

About this model

Overview

Upload a photo and get back a depth map. Three settings trade time for accuracy: the number of steps, how many predictions are averaged into the final map, and the working resolution, which follows your image unless you set a lower one.

What you can build

  • Depth guides for ControlNet and relighting work
  • Parallax animations from still photos
  • Depth passes for compositing and focus effects

Prompting tips

  • Keep the defaults for a first look; raise steps and the averaged predictions when edges look noisy.
  • Set a lower working resolution for large photos when you only need a rough map.

What to check

Check foreground separation and occluded edges. Do not interpret the image as measured distances.

Inputs

  • Image (required): Input image url.
  • Number of Inference Steps (optional): Number of denoising steps. Defaults to `10`. The higher the number, the more accurate the result, but the slower the inference. From 2 to 50.
  • Ensemble Size (optional): Number of predictions to average over. Defaults to `10`. The higher the number, the more accurate the result, but the slower the inference. From 2 to 50.
  • Processing Res (optional): Maximum processing resolution. Defaults `0` which means it uses the size of the input image. From 0 to 2048.

Output and limits

  • Small or dense text in the image can come out misspelled.
  • Hands, crowds and complex poses can show artifacts; generate a few variations and pick the best.
  • Some prompts or outputs can be refused by safety checks.

Price guide

  • up to: $0.28 / run