Exploring product rendering generation design catering to multi-emotional needs through the Superiority Chart-Entropy Weight method and Stable Diffusion model

The experience economy has shifted user demands towards emotionalization, emphasizing multi-emotional considerations as pivotal in design. This study addresses challenges in accurately determining emotional needs and the inadequacy of current intelligent design approaches. It proposes a method for designing multi-emotional product renderings by integrating the Superiority Chart-Entropy Weight method with the Stable Diffusion model within a big data framework. Initially, online user comments, hand-drawn sketches, and renderings of target products are collected. The Superiority Chart-Entropy Weight is then adopted to establish weights for multi-emotional needs, creating an allocation mechanism of these weights. Incorporating these multi-emotional weights, a Stable Diffusion model embedded with LoRa is trained to generate diverse rendering schemes. Finally, the Technique for Order Preference by Similarity to an Ideal Solution (TOPSIS) method is employed to select the optimal rendering scheme for 3D display. An experimental case study focusing on new energy vehicle renderings demonstrates the efficiency of this approach in precisely meeting users’ multi-emotional needs, thereby enhancing design efficiency and quality. Comparative experiments indicate that the method proposed in this study offers advantages in creating multi-emotional renderings. This study innovatively introduces a finer-grained multi-emotional needs confirmation method for users, overcoming the ambiguity and uncertainty of traditional recognition approaches, and develops a Stable Diffusion generation method tailored for product renderings, providing practical value in streamlining the conventional product design representation cycle and enhancing design efficiency, quality and user satisfaction.