Depth Map Generator: what it does
A depth map is a greyscale picture of distance: the nearer something is to the camera, the lighter it is (or darker, if you flip it). This generator estimates the depth of an ordinary photo with Depth Anything V2 Small, an open-source AI model (Apache-2.0) that runs in your browser on your own graphics card or processor. The model downloads once (27 MB) and is saved in your browser, so your photo is never uploaded. Choose grey or a colour map (Inferno, Turbo, Viridis), flip near and far, adjust the contrast, and download the depth map as a PNG at your photo’s size. Use it for 3D photo and parallax effects, depth of field and lens blur, displacement in 3D software, or as a depth guide for other image tools.
How to use it
- Drop a photo into the box, paste it with Ctrl+V, or tap to choose one from your phone.
- The first time, allow the one-time AI download. It is saved in your browser for next time.
- Wait while the AI measures depth on your device (about 10 seconds on our 4-core test PC), then pick grey or a colour map.
- Flip near and far or change the contrast if you need to, and download the PNG.
Tips
- Photos with a clear foreground and background (a person in a room, a product on a table, a street) give the most striking depth maps.
- The depth is relative, not in metres: it shows what is nearer and farther within this photo, so it can’t measure real distances.
- Most 3D and blur tools expect near = white. Flip it if your software wants near = black.
- Mirrors, glass and reflections confuse every depth AI: the reflection is often read as being far away.
- Colour maps such as Turbo make small depth differences easier to see; use grey for software.
Frequently asked questions
How do I make a depth map from a photo?
Add the photo and allow the one-time AI download. The depth map takes about 10 seconds on our 4-core test PC, less with a fast graphics card. Pick grey or a colour map, flip near and far if needed, and click Download PNG.
Is my photo uploaded?
No. The AI model is downloaded to your browser and runs on your device, so the photo never leaves it. After the first run you can switch on airplane mode and make more depth maps.
Which AI model is used?
Depth Anything V2 Small, an open-source monocular depth model released under the Apache-2.0 licence. We ship it with 8-bit weights to make the download about four times smaller; its output matches the original model almost exactly (correlation 0.9998 in our tests).
What resolution is the depth map?
The AI measures depth at about 518 pixels on the photo’s short side (392 on phones, to stay fast), and the depth map is then smoothly scaled up to your photo’s full size (up to 24 megapixels) for the download.
Is the depth map 16-bit?
No, it is an 8-bit PNG (256 levels of grey), which browsers can save directly. That is enough for most 3D photo, blur and displacement uses.