Related Experiment Video
Updated: Jun 1, 2026

Generation of Warfighter Avatars from Weapon Training Scene Images for Blast Exposure Simulations
Published on: December 6, 2024
AI image generation technology for visual communication design
Guohao Kong1, Long Zhang2, Junghyen Kim3
1The Graduate School of Advanced Imaging Science and film, Chung-Ang University, Seoul, 06974, South Korea.
This study enhances artificial intelligence (AI) image generation for visual design by improving requirement alignment and style control. The novel framework significantly reduces design cycles by over 91% while maintaining output quality.
Area of Science:
- Computer Vision
- Artificial Intelligence
- Graphic Design
Background:
- Current AI image generation struggles with design-specific requirements and consistent style control.
- Integrating AI into professional design workflows requires addressing alignment and controllability challenges.
Purpose of the Study:
- To systematically explore and enhance AI image generation for visual communication design.
- To develop an integrated framework for AI-driven design workflows.
- To establish evaluation metrics for AI-generated designs.
Main Methods:
- A novel framework combining diffusion models, CLIP-based cross-modal alignment, and ControlNet spatial constraints.
- Development of multi-dimensional evaluation metrics including generation quality and requirement matching.
- Experiments across posters, brand logos, and mobile interfaces using public and custom datasets.
Main Results:
- The integrated framework significantly reduces the design cycle by over 91%.
- The model demonstrates reliable generation under textual requirement noise and data distribution shifts.
- Mobile interface design scenarios showed the best performance.
Conclusions:
- The developed AI framework effectively addresses key challenges in visual design generation.
- Optimization paths are proposed for enhancing AI capabilities in design, including CLIP and ControlNet improvements.
- Limitations in data adaptability and dynamic generation are identified, guiding future research.
Related Concept Videos
Visual System
Once through the pupil, the light passes through the lens, a...
Design Example
Design Consideration
The factor of safety is another key aspect...
Managing Impressions
Design Example: Aggregate Gradation
The grading, or particle-size distribution, of sand is determined using sieve analysis, with standard sizes ranging from 150 μm to 10 mm (ASTM No. 100 sieve to 3⁄8 in. sieve). Sand is sampled...
Impression Management Techniques III: Aligning Actions
