Jan 24, 2025
Rendero Segmentation

Rendero Segmentation: How to Influence Outputs Using Segmentation Maps
Rendero is a cutting-edge visualization tool designed to transform the way architects approach rendering projects. The process begins with a base image that establishes the object's shape and details. From this starting point, a suite of advanced tools is at your disposal to enhance and refine your visualization. This article delves into the intricacies of each step, with a special emphasis on segmentation maps – the most powerful feature in Rendero AI, enabling unparalleled precision and creativity.
Base Image: The Key to Success
The first and most crucial step is choosing the base image. This represents your input visual data—whether a render, sketch, screenshot, or another image. The base serves as the foundation from which the entire composition develops, and its layout directly impacts the final result.
When creating the base image in Rendero, it’s essential to establish a clear structure and focus on the details you want to preserve in the final visualization. If certain minor details are not yet decided, you can allow the AI to fill them in. However, for well-defined design elements, we recommend incorporating these into the base image. For example, if you desire a specific texture in a particular area of the object, uploading an image with the texture mapped will help the AI accurately determine the scale and placement of the material. Additionally, smaller details may not be recognized in low-quality images. Therefore, we encourage uploading high-quality images to ensure the AI captures all fine details of your design.
Reference Image: Creating Atmosphere
The next layer in the process is the reference image. This element defines the “mood” of the result—its color scheme, lighting, and overall atmosphere can be directly derived from the reference image. For example, if you want your project to reflect modern minimalism with cooler tones, choose a reference that emphasizes these aspects. This step ensures harmony between technical components and visual aesthetics.
Text Specification: Precise Definition
The third level of creation allows you to define details through text, a highly effective method for precise specifications. There are two ways to write a prompt: the default pre-written style and the custom style.
In the default pre-written style, all you have to do is select the season, time of day, and fill out the text field describing the objects you want to include in your rendering—concrete walls, overcast sky, flowery garden, modern frameless windows. Textual descriptions give you detailed control over every aspect of the result and are excellent for emphasizing key project features.
The custom style offers even more detail and control. You will be greeted with two text fields, one positive and one negative, allowing you to fine-tune your prompts. This method requires a bit more work and adjustment but provides greater precision.
It's important to note that the prompt is not a chatbot. Giving it orders or asking it to perform special tasks like “make the wall behind the sofa pink” will not work. Instead, the prompt operates in a simpler, more precise manner—just describe what you want to see.
Segmentation Maps: Detailed Control Over Individual Elements
One of the most advanced features of Rendero AI is its ability to work with segmentation maps. These tools allow you to define individual segments of your image and influence how they are rendered. Segments represent different parts of the image. For example, in a rendering of an exterior house, you might separate the image into windows, walls, roof, and driveway. Conversely, an interior rendering of a living room might require more segments: walls, sofa, table, windows into the garden, hanging pictures, carpet, chairs, plants, and so on.
Currently, Rendero supports up to 16 segments, with the division being entirely up to the user. If you want to use segmentation to define each and every material, you can do so. Alternatively, if you prefer to select more general areas and leave the finer details for the AI to decide, that’s also an option.
Segmentation lets you control each section independently, which is particularly useful for complex visualizations or when precise adjustments to color and texture are required.
Manual Segmentation
Segmentation can be performed manually using simple tools, allowing you to outline specific areas, such as windows or other key features. Although this method is a little more time-intensive, it offers a simpler way for a little less skilled users.
Automatically Generated Maps
A more efficient approach involves using automatically generated segmentation maps, which can be created in any 3D software like 3Ds Max, Maya, Cinema 4D, or Blender. These maps are color-coded images where each material or object in the render is assigned a distinct color. This process is fast, effective, and allows for seamless integration of segmentation data into Rendero AI.

This color-coded image is the segmentation map.
Each color represents a specific material. Rendero automatically recognizes each shade as an individual segment.

Each segment allows you to specify the material or color you want to see in that area. There are a couple of key points to note: firstly, segments are not in-painters (yet) and cannot add new features or objects that are not present in the base image. Secondly, while segments do not influence each other directly, there can be indirect changes between them. For example, if you create a segment for the sky and prompt it as “overcast sky,” it will affect the entire image since the AI assumes the image was shot without direct sunlight. On the other hand, if you change the color of a small object, like a planter or a piece of furniture, you might not see any changes other than the color of the specific small object.
Segmentation Maps: Detailed Control Over Individual Elements
One of the standout features of Rendero AI is its ability to work with segmentation maps. These maps provide an advanced level of control over visualizations by breaking the image into distinct components, or segments, that can be individually manipulated. This approach is particularly valuable for complex visualizations, as it allows creators to focus on refining specific parts of the composition without affecting the whole.
What Are Segments?
Segments represent distinct parts of an image. For example, in the visualization of a building, segments could include:
Windows: The glass, frames, and shutters.
Frames: Borders or structural supports.
Terraces: Including flooring materials, railings, and any decorative elements.
Shading: Such as awnings, blinds, or overhangs.
Plaster: The building’s walls or facade, along with textural variations.
Roofing: Including tiles, slates, or other covering materials.
Each segment can be treated as a unique layer within the image, providing granular control over its color, texture, and overall appearance.
The Advantages of Segmentation Maps
Segmentation maps unlock new creative possibilities for users by enabling:
Precision Editing: Artists and designers can independently modify specific sections of an image without disturbing other elements.
Customization: Detailed text inputs for each segment allow for a highly tailored output, such as defining textures, materials, or lighting specific to a segment.
Time-Saving Automation: Automatically generated segmentation maps streamline workflows, especially for projects with extensive detail.
High-Quality Outputs: By focusing on individual elements, the resulting visualization achieves a professional and polished look, free from unintended inconsistencies.
Manual Segmentation: Maximum Detail Control
Manual segmentation involves using tools to manually define and outline each segment of the image.
How It Works: Users interact with the base image, drawing precise boundaries around elements like windows or doors. For instance, a user might trace the frame of a window or outline the roofline to define these as independent segments.
Use Case: This method is ideal for projects that require a high degree of customization, where automated tools may not recognize or define segments accurately.
Challenges: Manual segmentation can be time-consuming, particularly for large or highly detailed images. However, the tradeoff is unparalleled control over every aspect of the visualization.
Automatically Generated Segmentation Maps: Efficiency at Scale
For projects where speed and scalability are crucial, automatically generated segmentation maps are the preferred approach. These maps are typically created in 3D software like 3Ds Max, Maya, Cinema 4D, or Blender.
How It Works:
The 3D program generates a color-coded image, where each distinct material or component is assigned a unique color.
These colors correspond to different materials in the 3D model, such as concrete, glass, or wood.
Rendero AI then reads this segmentation map, interpreting the colors as separate segments.
Benefits:
Speed: Automating the segmentation process drastically reduces preparation time.
Accuracy: Modern 3D programs often produce highly accurate segmentation maps, minimizing the need for corrections.
Integration: These maps integrate seamlessly into Rendero AI, enabling creators to refine their visualizations further.
Exporting material ID maps from 3D software like 3Ds Max, Maya, Cinema 4D, or Blender is a crucial step in creating segmentation maps for Rendero AI. Here's a step-by-step guide for each software:
1. 3Ds Max
Material ID maps in 3Ds Max are generated using the Render Elements feature.
Steps:
Assign Material IDs:
Open the Material Editor.
Assign unique Material IDs to different materials in your scene. You can find this option under the material properties.
Set Up Render Elements:
Open the Render Setup window (F10).
Go to the Render Elements tab.
Click Add and select VRayMtlID (for V-Ray users) or Material ID for default renderer users.
Ensure that the Render Elements include a Material ID pass.
Render the Material ID Map:
Render the scene as usual. The Material ID pass will output a color-coded map where each material ID is represented by a distinct color.
Save the Material ID Map:
Save the output map (usually in formats like PNG or EXR for better quality).
2. Maya
Maya supports material ID maps through render layers or AOVs (Arbitrary Output Variables) in Arnold.
Steps:
Assign Material IDs:
Open the Hypershade.
Assign different Shaders or use the aiUtility shader for generating flat colors for different objects or materials.
Set the Color Mode of the aiUtility shader to Object ID or Material ID.
Set Up Render Layers:
Create a new Render Layer for the material ID pass.
Assign objects with different materials to separate layers if necessary.
Enable AOVs in Arnold:
Open the Render Settings.
Go to the AOVs tab.
Add a custom AOV like cryptoMaterial or use existing options for material-based segmentation.
Render the Material ID Map:
Render the scene using Arnold, ensuring the AOV outputs include your material IDs.
Save the Map:
Save the rendered material ID map as an image file.
3. Cinema 4D
In Cinema 4D, material ID maps are typically created using multi-pass rendering.
Steps:
Assign Object Buffers or Material Tags:
Right-click on an object and add a Compositing Tag.
In the tag’s settings, enable Object Buffer and assign a unique buffer number.
Alternatively, assign different Materials with distinct colors.
Set Up Multi-Pass Rendering:
Open the Render Settings.
Go to the Multi-Pass tab.
Add a Material Color or Object Buffer pass, depending on your workflow.
Render the Multi-Pass Output:
Render the scene.
The Material Color pass will provide a color-coded material ID map.
Save the Output:
Save the material ID map from the multi-pass output files.
4. Blender
Blender uses ID Masks or Cryptomatte for creating segmentation maps.
Steps:
Enable Cryptomatte in Cycles:
Switch to the Cycles Renderer in the Render Properties.
In the View Layer Properties, enable Cryptomatte.
Select Object, Material, or Asset depending on the level of segmentation you need.
Assign Materials:
Ensure that each object or material in your scene has a distinct material assigned to it.
Render the Cryptomatte Pass:
Render the scene. The Cryptomatte pass will generate a segmentation map with distinct colors for different materials.
Extract the Cryptomatte Data:
Open the Compositor.
Add a Cryptomatte Node.
Connect it to the Render Layer and use the Picker tool to isolate specific material IDs.
Save the Material ID Map:
Save the output from the compositor as an image file.
Tips for All Software:
Use lossless file formats like PNG or EXR to ensure the segmentation map maintains its color fidelity.
Test your segmentation maps in Rendero AI before proceeding with further adjustments to confirm that all segments are correctly identified.
Use distinct colors for different materials to avoid confusion or blending issues during processing.

This color-coded image is the segmentation map.
Each color represents a specific material. Rendero automatically recognizes each shade as an individual segment.
Enhancing Segments Through Textual Definitions
Once segments are defined, whether manually or automatically, Rendero AI allows users to add descriptive details to each segment.
Examples of Textual Inputs:
"Light concrete, slightly worn, with pronounced texture in areas of material junctions."
"Clear glass with a faint reflective sheen, emphasizing transparency."
"Polished wooden flooring, warm tones, minimal grain visible."
Purpose of Textual Definitions:
These inputs act as guidelines for Rendero AI to enhance the segment's rendering.
They provide creative direction, ensuring that the final visualization aligns with the user’s vision.

You can then add a text description to each segment, such as: “light concrete, slightly worn, with pronounced texture at material junctions,” and so on.


This is the final visualization output.

When overlaid with the base image, you can see that elements such as windows, roofs, and water levels have not shifted.
Conclusion
Segmentation maps represent the pinnacle of control in Rendero AI.
These tools provide precise, detailed manipulation of individual image elements, enabling you to achieve professional, high-resolution results. When used in conjunction with base images, references, and text specifications, segmentation maps become an incredibly powerful resource, propelling your visualizations to new heights.
Whether you’re an architect, designer, or artist, Rendero delivers the flexibility and precision necessary to bring your creative visions to life.
This is the final visualization output.
When overlaid with the base image, you can see that elements such as windows, roofs, and water levels have not shifted.
✻
✻
✻