Transform Your Designs with AI-Generated Realistic Renders

Table of Contents

1. Introduction
2. Creating a Mood Board
   • Using Style Source Book
   • Selecting Furniture and Textures
   • Adding Basic Elements
   • Testing with Journey Bot
3. Remodeling with Room GPT
   • Choosing a Room Style
   • Uploading the Image
   • Exploring Different Options
4. Using Stable Division
   • Using Stable Diffusion Image-to-Image
   • Using ControlNet Extension
   • Comparing Results with Mid Journey
5. SketchUp Modeling
   • Creating a Basic Room Scene
   • Adding Minimal Details
   • Testing with Journey Bot and Room GPT
6. Comparing Results
   • Overall Image Quality Comparison
   • Comparison of 3D Model Results
7. Conclusion

Turning a 3D Model into a Realistic View

Today, I will be exploring the process of turning a simple 3D model into a realistic view. In a previous video, I experimented with creating realistic renders from HAND drawings. This time, I want to see if I can achieve a similar result using a basic 3D model and reference images.

Creating a Mood Board

To start, I decided to Create a mood board for an interior living room scene. I used a Website called Style Source Book, which made it easy to find and incorporate furniture products and textures. I kept the mood board simple to test the capabilities of AI Tools for interior design and architecture.

Using Style Source Book, I selected a nice sofa for the living room and added a couple of plants and an abstract painting on the wall. I also experimented with adding a wallpaper to one section of the background. Lastly, I added a simple wooden flooring to complete the scene. Please note that the quality of the mood boards may not be perfect as my focus was on testing the AI tools.

Next, I wanted to test the capabilities of the AI Tool called Journey Bot. I uploaded my mood board and described the materials, lighting, and other important details using basic keywords. I also adjusted the image weight to see how it would affect the results. I generated multiple versions to compare and analyze later.

Remodeling with Room GPT

Another tool I wanted to explore was Room GPT. With this website, I could choose a specific style for my space and upload the image. The website would then remodel the image Based on the chosen style. I started by using the minimal style option and obtained a redesigned image.

Room GPT provides three free options to test out the website, with the option to upgrade to a premium version for more features. I also experimented with other styles, such as modern and tropical. Though I noticed similarities to stable division in the results, the options were limited as I could only choose a room style without making further modifications.

Additionally, I tested stable diffusion image-to-image version with the help of the ControlNet extension. Following a similar workflow as my previous video for turning hand drawings into realistic views, I described my image to stable division and generated the results. I compared the outcomes to see the differences and similarities.

SketchUp Modeling

For the Second part of the video, I used a SketchUp modeling tool to create a basic room scene. I extruded walls, added a window, and positioned a simple sofa in the space. I also placed a rectangular box in the corner to serve as a plant. To create some lighting effects, I turned on the lights, which were coming from the window. Lastly, I added a circular table in the middle and painted different surfaces with different colors for reference.

To transform the SketchUp model into a realistic view, I employed the same process as before. First, I used Journey Bot, describing the space and materials. Then, I tested various image weights and generated multiple versions to compare the results. Next, I used Room GPT to remodel the room scene into different styles, exploring different options.

Comparing Results

After testing all the tools and techniques, it's time to compare the results. When comparing the mood board results, Journey Bot produced images with overall good quality. However, it creates new versions inspired by the uploaded images instead of using the images directly. Room GPT allowed me to choose different styles for the mood board, but it lacked control over the process. Stable division generated results closer to the original image but had limitations in understanding lighting and certain elements.

Moving on to the 3D model results, I was highly impressed with the outcomes achieved through Journey Bot. It accurately captured all the elements in the uploaded image, including windows and the box used as a plant. Room GPT, on the other hand, provided a convenient way to generate different styles for the 3D model. Lastly, stable division performed decently but couldn't match the quality of Journey Bot.

Conclusion

In conclusion, the process of turning a 3D model into a realistic view involves using AI tools like Journey Bot, Room GPT, and stable division. Each tool offers its unique capabilities and limitations. Journey Bot excels in creating realistic views with good overall image quality. Room GPT allows for exploring different styles easily but lacks control. Stable division provides results closer to the original image but may struggle with certain elements. By comparing the results, it becomes evident that Journey Bot offers the most satisfying outcomes.