The ComfyUI node interface plays a key role. It allows you to create and manage complex image processing scenarios, avoiding the illogicality and unexpected effects that often occur when using other tools. This allows you to customize each generation saudi arabia phone number stage and provide maximum flexibility and control over the final result.
I as the main tool for working with images allowed us to significantly improve the quality and predictability of the resulting visualizations, which is especially important for the implementation of complex hotel projects.
An important condition of the task was to find forms of bionic facades with a clear emphasis on the floors and the possibility of their architectural play. This requirement became the basis for creating a prompt, which was further expanded using ChatGPT. A rough description of the task was issued in ChatGPT, and as a result, a specific request for image generation was received. The request was then used in Stable Diffusion.
It is necessary to add that ChatGPT “remembers the dialogue” and it can be asked to change its own answer in different parts of it, so you can improve the prompt an infinite number of times without effort.
As a result, the process of creating visualizations for the hotel project looked like this:
Specification of the task : creation of bionic facades with an emphasis on ceilings
Formulating a Prompt : Describing the Task in General Terms
Extending Prompt with ChatGPT : Refine Query to Generate Image
Using Stable Diffusion : Generating images on demand
This approach allowed us to precisely formulate the requirements and obtain high-quality visualizations that corresponded to the original concept.
Fig. 7. Initial query for ChatGPT
Artificial Intelligence in Design
Fig. 8 - 12. Examples of images generated by Stable Diffusion
Artificial Intelligence in Design
Artificial Intelligence in Design Artificial Intelligence in Design Artificial Intelligence in Design Artificial Intelligence in Design
Case 2. Preprocessing images to obtain more detailed visualizations. Testing the AI+AI bundle
To create sketches of building facades with maximum detail, simple forms from Revit were processed in the Kandinsky service. This tool edited the edges of Revit forms and added details. The resulting image was then used in Stable Diffusion.
Fig.13. Original model in Revit
Artificial Intelligence in Design
Fig.14. Textures and landscaping added in D5
Artificial Intelligence in Design
Fig.15. Image detailing by the Kandinsky service
Artificial Intelligence in Design
Fig.16 - 18. Generating Stable Diffusion visualizations
Artificial Intelligence in Design
Artificial Intelligence in Design Artificial Intelligence in Design
The process of creating visualizations for a hotel project in the AI+AI bundle, step by step:
Original Image . Creating a Base Image in Revit
Pre-processing in Kandinsky . Running the base image, adding details and outlines of forms
Using in Stable Diffusion . Further generation of visualizations and their refinement
Conclusion: using the AI+AI combination allows you to improve the quality of the initial image and get a more detailed and accurate result at the final stage of processing.
In subsequent iterations, the resulting images were passed through a resolution magnifier, also written on the basis of Stable Diffusion and ComfyUI. This tool magnifies the original image by a specified number of times. In this case, an image with a resolution of 1280×1264 was enlarged to 5120×5056. Important: the resolution magnifier did not add additional details and noise, the image details are fully preserved.
Case 3. Visualizations of individual buildings at a hotel facility
The approaches and order of actions when working with building visualizations are the same as with the areas described above. The work of the tools is shown in the figures.
Fig. 19. Original building model from Revit
Artificial Intelligence in Design
Fig. 20 - 23. Generating visualizations of the Stable Diffusion building
Artificial Intelligence in Design
Artificial Intelligence in Design Artificial Intelligence in Design Artificial Intelligence in Design
In conclusion, it should be noted: when developing hotel projects, the use of artificial intelligence increases the number of fore-sketches and concepts by 5 or more times! AI significantly reduces labor-intensive processes at the earliest and most volatile stage of design.
For further development of the project, other AI tools can be connected. In particular, the Architectures service allows you to quickly obtain a developed model with the possibility of fine control. It uses advanced algorithms to automate and optimize the development of architectural projects.
Of course, none of the artificial intelligence tools today are capable of replacing humans in complex technical processes, but when used creatively, they significantly reduce the labor hours spent on routine tasks and vastly expand creative horizons.