Product Innovation Design Method Based on BP Neural Network

Innovative product design is the core problem of modern industry and source power enterprise development. With continuous improvement China’s material level, innovation has changed from functional needs to emotional needs, method centered on users’ attracted much attention. emergence new technologies such as artificial intelligence big data, automation advanced algorithms innovative methods for online decision driven by multimodal multimedia data in e-commerce have become inevitable trends field design, it a challenge improve ability products using information technology. In order solve this problem, based research BP neural network technology, paper puts forward an image-Kempi method. The color features images are reconstructed integrated, then transferred modeling, so generate with both content image modeling style features, which brings visual inspiration users. results show that KENPI combines theories Kansei engineering, convolutional network, transfer, establishes mapping model between elements semantics. Convolutional transfer used extract, reconstruct, integrate diagrams, them images. By evaluating quality comparing semantics before after migration, validity feasibility migration verified. Based nonlinear attribute space semantic constructed, generalization evaluated, verifies effectiveness important theoretical guiding significance improving enterprises, enhancing competitiveness, customer satisfaction.