Image To Relief

Image To Relief

Turns an uploaded picture into a 3D relief STL file by estimating depth, and lets you reshape the depth map with a curve editor before you export. Suits makers who want a height model from a photo.

Made by oedevs on Hugging Face. These notes are Mindket's, written from its Space page and README; the app's own files and license are on Hugging Face.

About

What it is: Image To Relief turns a picture you upload into an STL file of a relief. It asks the Depth Anything V2 model to estimate depth, then gives you a curve editor and texture controls so you can adjust the depth map before anything is exported.

Who it suits: It suits anyone who wants a height-based 3D model from a flat image and wants a say in how the depth looks.

Title: Who might use this app
Industry | Who uses it | How they use it | Example
3D printing | Makers and hobbyists | Turn a photo into a model file that has height | Upload a photo and export an STL
Design | Designers | Reshape depth so the subject stands out from the background | Apply the foreground emphasis preset
Computer vision | Developers | See how a depth estimate becomes a height map | Edit the curve and watch the previewed height map change
Local tinkering | Anyone with a spare computer | Raise the pixel limit for larger images | Start the app with a higher limit

If you need to edit a mesh by hand or sculpt freely, you can skip it.

Where to use it: You can use it in a browser, in Docker, or with plain Python.

• Web browser: the Space's page on Hugging Face, with nothing to install
• Windows, macOS or Linux with Docker: a local copy built from the Dockerfile
• Python on your own computer: run app.py after installing the requirements

Title: What the app is good for
Use case | Who | What they get back
Make a relief from a photo | Makers | An STL file built from the adjusted depth
Fix a flat or harsh depth estimate | Designers | A reshaped depth map with the curve and presets
Work on bigger images | Local users | A higher processing limit than the hosted Space allows

How to set it up: Pick one of two routes. The hosted Space needs nothing. For a local copy you need the items below.

• Docker for the container route, or Python with pip for the plain route
• pip, to fetch the Hugging Face command line tool

1. Install the huggingface_hub package.
2. Download the Space's files into a folder.
3. Move into that folder.
4. Build the Docker image.
5. Run the container and publish port 7860, then open that port in your browser.

pip install -U huggingface_hub
hf download oedevs/image-to-relief-grayscale --repo-type space --local-dir image-to-relief-grayscale
cd image-to-relief-grayscale
docker build -t image-to-relief-grayscale .
docker run -it -p 7860:7860 image-to-relief-grayscale

Without Docker, install the requirements and start the app directly.

pip install -r requirements.txt
python app.py

How to use it: Work through the page from top to bottom.

1. Add your picture, then press Estimate Depth Map.
2. Reshape the depth values in the histogram and curves editor.
3. Change the texture strength and smoothing settings and watch the height map preview react.
4. Check the previewed height map, set the STL dimensions and click Generate STL from Adjusted Depth.

Title: The controls you will use
Tool | What it does
Estimate Depth Map | Runs the depth model on the image you uploaded
Depth histogram + curves editor | Remaps depth values when you drag, add or remove curve points
Curve presets | Apply midtone contrast, foreground emphasis or background compression
Luminance Texture Strength | Adds surface texture taken from the image's brightness
Generate STL from Adjusted Depth | Builds the STL file from the depth map you tuned

To raise the processing limit on your own computer, pass a flag when you launch.

python app.py --max-resolution-pixels 4000000

Example: You upload a photo, click Estimate Depth Map, then drag a point on the curve so the main subject sits higher than the background. After you click Generate STL from Adjusted Depth, you get an STL file of that relief.

For newcomers and experts: If you are new to this, open the Space, upload any image and click Estimate Depth Map. Look at the result before you touch the curve, then try one preset.

If you work with depth models, the app uses Depth Anything V2 Large for the estimate and lets you remap its values by hand. The texture is scaled to the current height range, so a larger Max Z-Height does not make the surface texture twice as deep. HEIC and HEIF uploads are decoded on the server with pillow-heif.

Upload an image
Estimate the depth map
Reshape the curve and texture
Preview the height map
Generate the STL file

Good to know: The hosted Space has a cap you cannot change, and a local copy can raise it.

• Hosted limit: processing is capped at 1,500,000 pixels so shared users cannot overload the Space
• Local limit: raise it with the flag above, or with the DEFAULT_MAX_RESOLUTION and MAX_RESOLUTION_LIMIT environment variables
• HEIC: if a browser cannot preview a HEIC file, the upload path should still work
• Versions: the package versions in requirements.txt are not pinned yet
• License: mit
• Cost to run: none named
• Data it reaches: the image you upload, which the app processes with the Depth Anything V2 Large model
• Not for: raising the pixel cap on the hosted Space, because the web page does not offer that control

On Hugging Face, 13 November 2026

• Likes: 3
• Built with: docker
• License: mit

Links

• Space: https://huggingface.co/spaces/oedevs/image-to-relief-grayscale
• Files: https://huggingface.co/spaces/oedevs/image-to-relief-grayscale/tree/main

Try it

Open the Space in your browser, or copy it to your own Hugging Face account from its page. Nothing in this kit runs by itself.