> For the complete documentation index, see [llms.txt](https://fater.gitbook.io/fater-ai-docs/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://fater.gitbook.io/fater-ai-docs/generating-images-with-ai/preprocessing-for-ai-edges-and-depth-maps.md).

# Preprocessing for AI: Edges and Depth Maps

To effectively use **Controlled AI Models** that rely on Edges or Depth maps, Fater provides tools within the **Image Editor** to generate these specialized images directly from your existing canvas content. These "preprocessed" images are then typically used as **Control Layers**.

These tools are found on the **Inpaint Toolbar**.

***

### 1. Generating an Edges Map

An Edges map highlights the outlines and significant contours of your image, providing structural guidance to AI models.

* **How to Generate:**
  1. **Prepare Canvas & Generation Area:** Ensure the visual content from which you want to extract edges is visible on your **Inpaint Canvas** and falls within the **Generation Area Bounding Box**. The tool will process what's inside this box.
  2. **Access Options (Optional):** To adjust sensitivity, **right-click** the **Generate Edges Map** (CPU 🤖 icon) on the Inpaint Toolbar. A popover will appear:
     * **Low Threshold:** Affects how readily weaker edges are detected.
     * **High Threshold:** Determines strong edges that are definitely included.
     * Adjust these values as needed.
  3. **Generate:** Click the **Generate Edges Map** (CPU 🤖 icon) on the Inpaint Toolbar.
* **Result:** A new layer is created and added to your **Layers Tab**. This layer will be a black and white image, with white lines representing the detected edges. It's often named something like "Edges: \[thresholds]".
* **Next Steps:** You can then designate this new layer as a Control Layer and set its "Preprocessed Type" to `edges`. *(See* [Working with Control Layers](/fater-ai-docs/working-with-layers/working-with-control-layers.md)*)*

{% hint style="info" %}
This operation doesn't involve any model, and so doesn't cost any credit.
{% endhint %}

***

### 2. Generating a Depth Map

A Depth map represents the perceived distance of objects in your image from the viewer, providing 3D spatial information to AI models.

* **How to Generate:**
  1. **Prepare Canvas & Generation Area:** Ensure the visual content for depth analysis is visible on your **Inpaint Canvas** and within the **Generation Area Bounding Box**.
  2. **Access Options (Optional):** To change generation type or tiling, **right-click** the **Generate Depth Map** (Mountain ⛰️ icon) on the Inpaint Toolbar. A popover will appear:
     * **Depth Generation Type:**
       * **Monotile:** Processes the area as a single pass. Faster, but may have less detail for complex scenes.
       * **Multitiles:** Divides the area into smaller tiles for processing, then combines them. Can produce more detail but is significantly slower (may take up to 5 minutes).
     * **X/Y Split Tiles (for Multitiles):** If "Multitiles" is selected, define how many tiles to split the area into horizontally (X) and vertically (Y).
     * Adjust these settings as needed.
  3. **Generate:** Click the **Generate Depth Map** (Mountain ⛰️ icon) on the Inpaint Toolbar.
* **Result:** A new grayscale layer is created and added to your **Layers Tab**. Lighter areas represent closer elements, darker areas represent further elements. It's often named "Depth Map".
* **Next Steps:** Designate this new layer as a Control Layer and set its "Preprocessed Type" to `depth`. *(See* [Working with Control Layers](/fater-ai-docs/working-with-layers/working-with-control-layers.md)*)*

***

**Important Considerations:**

* **Generation Area is Key:** Both tools operate strictly on the content visible within the current **Generation Area Bounding Box**.
* **Visible Layers:** The tools analyze the composite image formed by all currently *visible* layers within that bounding box.
* **New Layer Output:** The generated Edges or Depth map is always created as a **new control layer**, leaving your original layers untouched.

These preprocessing tools are essential for preparing the visual guidance needed by Fater's advanced "Controlled" AI models, allowing for more precise and structured image generation.