Analyzing the packaging design evaluation based on image emotion perception computing.

Area of Science:

• Human-Computer Interaction
• Artificial Intelligence
• Consumer Psychology

Background:

• Modern packaging design increasingly focuses on functionality and individual consumer needs.
• Conventional design methods rely on subjective designer experience and struggle with emotional analysis.
• Previous research on emotional image analysis often fails to capture individual viewer responses.

Purpose of the Study:

• To develop an objective and dynamic approach for evaluating packaging design based on emotional impact.
• To improve the user-product interaction lifecycle through enhanced sensory experiences.
• To overcome limitations of traditional design evaluation by incorporating computational emotion perception.

Main Methods:

• Proposed a packaging design evaluation approach based on image emotion perception computing (PDE-IEPC).
• Utilized a deep Long Short-Term Memory (LSTM) model integrated with emotion perception technology.
• Employed the Dynamic Multi-task Hypergraph Learning (DMHL) approach within Emotion Perception Computing, considering graphical, social, spatial, and locational data.
• Leveraged the Image-Emotion-Social-Net dataset, comprising over 1 million images from Flickr, for personalized emotion categorization.

Main Results:

• The PDE-IEPC approach demonstrated superior performance in personalized emotion categorization compared to existing methods.
• Achieved a high packaging design quality rate of 94.1%.
• Attained a performance success rate of 97.5% with a mean square error rate of 2%.

Conclusions:

• The developed PDE-IEPC method offers an effective and dynamic evaluation of packaging designs based on emotional impact.
• The approach provides a more personalized and accurate understanding of user emotions in product interaction.
• This computational method enhances packaging design quality and user experience, addressing limitations of traditional subjective evaluations.