Microsoft Research Asia and Peking University develop COLE, an AI tool that generates high-quality graphic designs with editable text and visual elements.
Graphic designers and those who rely on them may soon face a game-changing disruption in their profession. COLE, an AI tool developed by researchers at Microsoft Research Asia and Peking University, has the potential to revolutionize graphic design. By combining various AI models and the open-source graphics renderer Skia, COLE can generate not only visually appealing images but also include text that seamlessly integrates with the design. This breakthrough in AI-generated graphic design could have far-reaching implications for the industry.
The Complex Framework Behind COLE
The development of COLE involved combining different AI models, including fine-tuned versions of Meta’s Llama2-13B, DeepFloyd IF, LLaVA1.5-13B, and GPT-4V. The researchers also trained the AI on 100,000 high-quality raw graphic design images from the internet. This framework allows COLE to generate crisp and organized graphic designs by extracting the background layer and describing it in text.
Auto-Generated Designs with Editable Elements
One of the most impressive features of COLE is its ability to produce images with distinct editable blocks for text and objects within the design. This means that users can modify the text, font, or even introduce new visual elements within the COLE framework itself. The AI programs work together to create an image from scratch, and if the user is dissatisfied with the result, they can easily make changes without starting from scratch or using external software.
Competitive Results and Future Potential
The results obtained from COLE are highly competitive in terms of quality, even compared to other text-to-image generators like DALL-E 3. The researchers tested COLE on 200 different graphic design projects, including advertisements, event promotions, and marketing materials. While COLE has limitations, such as the inability to change the arrangement or placement of typography blocks and support for only one color of typography per image, the researchers plan to address these issues in future iterations.
Redefining the Role of Graphic Designers
COLE raises questions about the future of graphic design. While it may initially seem like a threat to the profession, the researchers designed COLE to allow users to refine the output and integrate human expertise when necessary. This suggests that graphic design training would still be valuable in getting the best results from the AI tool. However, COLE also empowers individuals without graphic design expertise to generate high-quality designs, potentially democratizing the field.
Conclusion:
COLE represents a significant breakthrough in AI-generated graphic design. Its ability to generate visually appealing designs with editable text and visual elements opens up new possibilities for both professionals and individuals without graphic design expertise. While COLE may not replace graphic designers, it has the potential to redefine their role and make high-quality design accessible to a broader audience. As COLE continues to evolve, it will be fascinating to see how this AI tool shapes the future of graphic design.
Leave a Reply