Adaptive advertisement delivery system and method based on emotion and environment perception
By using multimodal data analysis and simulating collaborative discussions among humanized agent clusters, personalized advertising material packages are generated. This addresses the lack of fine-grained understanding of the interaction between emotion and environment in existing advertising methods, enabling a more refined description of users' real-time psychological state and improving the accuracy of advertising decisions.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- BEIJING HONGTU XINDA TECH CO LTD
- Filing Date
- 2026-04-08
- Publication Date
- 2026-06-02
AI Technical Summary
Existing advertising methods struggle to capture the immediate psychological state of users at specific moments, influenced by both emotional fluctuations and environmental changes. They lack a granular understanding of the interaction between emotions and the environment, resulting in advertising decisions remaining at the level of relevance matching. Furthermore, existing emotion recognition methods lack a unified representation and in-depth collaborative analysis of visual, audio, behavioral, and environmental information, leading to insufficient accuracy.
By collecting multimodal user data and generating a unified embedded representation using a large-scale multimodal basic model, a simulated humanized agent cluster composed of emotion interpretation agents, environmental context agents, and multiple simulated agents with different personalities is activated. Through multiple rounds of collaborative discussions, the causal interaction between emotions and the environment is analyzed, and personalized advertising material packages with zero template dependence are generated and deployed in conjunction with a real-time bidding environment.
It enables a refined description of users' real-time psychological state, improves the accuracy of advertising decisions, makes advertising more in line with users' current state, and enhances the personalization and effectiveness of advertising.
Smart Images

Figure CN122134405A_ABST