An algorithm for optimizing psychological regulation strategies for college students based on image recognition and...

Area of Science:

• Artificial Intelligence
• Psychology
• Computer Science

Background:

• College students face increasing academic pressure and emotional challenges.
• Traditional psychological support methods are often subjective and lack real-time adaptability.
• There is a need for intelligent, adaptable systems for psychological management.

Purpose of the Study:

• To develop an algorithm for optimizing psychological regulation strategies using image recognition and reinforcement learning.
• To incorporate a Weighted Butterfly-Twin Delayed Deep Deterministic Policy Gradient (WB-TwinD3PG) Model for enhanced adaptive support.
• To provide personalized interventions for college students' mental well-being.

Main Methods:

• Utilized a dataset of ~10,000 psychological records.
• Applied pre-processing techniques: face detection, alignment, and illumination normalization.
• Employed Local Binary Patterns (LBP) for feature extraction and WB-TwinD3PG for reinforcement learning.
• Agent selected personalized interventions based on emotional state feedback.

Main Results:

• The WB-TwinD3PG model demonstrated superior performance over traditional methods.
• Achieved high scores in recall (0.921), F1-score (0.927), precision (0.934), and accuracy (0.913).
• The system provides adaptive, robust, and personalized mental health support.

Conclusions:

• The research presents a practical framework combining image-based emotion analysis and advanced reinforcement learning.
• This approach offers effective psychological well-being management for college students.
• The developed algorithm optimizes psychological regulation strategies for improved mental health outcomes.