Drop your image here, or click to browse
JPG · PNG · WebP · processed locally in your browser
Preview appears here after conversion
Upload an image and click Convert
✅ Result Ready
How Image to Depth Map Works
A depth map is a greyscale image where pixel brightness encodes distance from the camera. White pixels are close; black pixels are far. Depth maps are fundamental to a huge range of tasks: 3D compositing in After Effects, distance fog in game engines, point-cloud generation in scientific software, and the depth-of-field blur effect in portrait photography apps. This tool generates a depth map in two modes. Classic Mode converts your photo to greyscale luminance — a fast approximation that works well when the subject is bright against a dark background. AI Mode runs a MiDaS-class depth estimation model in your browser, producing a physically motivated depth estimate from appearance cues like perspective, relative size, and occlusion.
Both outputs are exported as standard PNG files at the full resolution of your source image. A 16-bit PNG option is included for AI Mode, which gives finer tonal gradations for use in professional VFX pipelines where high bit-depth matters.
Step-by-Step Guide
- Upload your image. Any JPG, PNG, or WebP photo. No specific requirements on subject matter.
- Select a mode. Classic for instant luminance-based results; AI for a scene-aware depth estimate.
- Choose output bit depth. 8-bit for general use and game engines; 16-bit for VFX compositing where precision matters.
- Click Convert. The depth map is generated and shown as a preview.
- Download the PNG. Import into After Effects, Blender, Unreal Engine, or any tool that accepts depth maps.
Best Use Cases
- Portrait depth-of-field blurs in Photoshop or Affinity Photo using the depth map as a blur mask.
- Game engine fog and distance effects: Use as a custom depth texture in Unity or Godot shaders.
- 3D photo effects in social media editors that accept depth maps for parallax or 3D pan effects.
- Input to other tools: Feed the depth map into our Image to Point Cloud tool for a photogrammetry-style preview.
Honest Limitations
- Not metrically accurate. The AI depth map estimates relative depth, not absolute distance in metres. It cannot replace a LiDAR sensor or stereo camera for measurement tasks.
- AI can misjudge reflective or transparent surfaces. Mirrors, glass windows, and water confuse the model's depth estimates. Classic Mode is sometimes more predictable for these cases.
- High-frequency detail may be smoothed. The AI model operates at 384×384 internally and upscales, so very fine edges may appear slightly blurred in the depth map.
Frequently Asked Questions
After Effects (Camera Lens Blur effect), Nuke (DepthOfField node), Blender (Defocus compositor node), Unity (depth-of-field post-processing), Unreal Engine (custom depth buffer), DaVinci Resolve (depth map compositing), and any tool supporting greyscale masks.
A depth map encodes distance from the camera as a single greyscale value. A normal map encodes the direction of a surface at each pixel as an RGB colour (red=X, green=Y, blue=Z). Depth maps are used for fog, blur, and parallax effects; normal maps are used for lighting and shading in game engines.
Related Tools
Written by the Image → 3D team · Last updated