Low-Carbon Design of Green Packaging Based on Deep Learning Perspective for Smart City

China's packaging business has started to exhibit a green, eco-friendly, and low-carbon development trend as a result of the increased attention being paid to environmental issues on a global scale. This paper aims to investigate the impact of deep learning model application in low-carbon package design from the viewpoint of smart cities. The low-carbon design of the flower and fruit tea packaging was used as an example here to study the low-carbon design mode and process of green packaging as well as the product's attributes and low-carbon green performance. The evaluation model for packaging design was then constructed based on the BP neural network algorithm training phases to assess the emotional worth of consumers' green packaging. The paper's findings demonstrated the viability of the BP neural network model, which had the best prediction performance in the 78th epoch. There is no difference between the model's projected and actual values, indicating the model's strong classification performance and capacity to create a relationship between color features and objective data and values associated with emotional evaluation. The findings of this paper offered favorable implications for achieving low-carbon green packaging and can be put into practice. They can also develop low-carbon design ideas for green packaging and reduce package pollution to the environment.