AdPercept: Visual Saliency and Attention Modeling in Ad 3D Design

Large language models (LLMs) have gained prominence in various fields, including marketing and advertisement design, where visual engagement plays a critical role. Traditional approaches in 3D ad design often overlook the nuanced elements that influence viewer attention. To address these challenges, we introduce AdPercept, a sophisticated framework that harnesses advanced computer vision techniques for modeling visual saliency and attention within 3D advertisements. AdPercept operates in two phases: first, it implements a saliency detection algorithm to pinpoint eye-catching regions, and second, it employs an attention model that mimics human attention patterns for optimizing advertisement layouts. This integration of psychological principles of visual perception enhances both aesthetic appeal and viewer engagement. Experimental results demonstrate that AdPercept outperforms conventional methodologies in capturing user attention and bolstering interactivity in ads.