SYSTEM AND METHOD FOR VISUAL CONTENT GENERATION AND ITERATION

Smart overview of the Invention

A system designed to enhance the creative process involves generating and iterating visual content with a focus on diverse design ideas. It utilizes a processor and memory that store instructions to create related and semantically diverse texts based on user prompts, employing a generative language model. Subsequently, these texts are used to generate corresponding images through a generative visual model.

Addressing Design Fixation

Design fixation, where users become overly focused on limited solutions, can hinder the creative process. To combat this, the system encourages imaginative exploration by maximizing the semantic diversity of AI-generated ideas. This approach allows users to explore new visual designs and refine existing ones without being constrained by familiar concepts.

Functionality of the System

The system operates in two primary modes: low-diversity and high-diversity. In low-diversity mode, it generates images directly based on user prompts. In high-diversity mode, it generates a set of semantically diverse texts from prompts and then creates images based on selected texts, fostering broader exploration of design ideas.

Refinement and Iteration

In addition to generating new ideas, the system supports refinement through iterative processes. It can take previously generated images as prompts to create related texts, which are then used to produce new images. This iterative approach allows users to enhance specific features of their designs effectively.

Advantages of the System

The disclosed system offers several advantages over traditional methods, including assistance in overcoming design fixation, providing diverse and inspirational content, and reducing the time and cost associated with generating new visual designs. The quick turnaround time enables users to efficiently iterate through ideas, facilitating a more dynamic creative process.