Intelligent Design of Home Products Based on Digital Intelligence Technology

Abstract

Abstract Aiming at the problems of high cost and too much reliance on the creator’s inspiration in traditional home product design, this paper combines digital technology to design home products. Specifically, the attention mechanism is utilized to improve the YOLOv5 network, and the home products are identified through the improved technique to obtain the home product dataset. The generative adversarial network is trained using the dataset, the new design scheme is output through the trained generative adversarial network, and the generative model is evaluated using the IoU metrics to realize the intelligent design of home products. After the model design is completed, the home products are intelligently designed according to the customer’s requirements. The IoU values of the 3D voxel model are 0.6858, 0.6872, 0.7834, and 0.6268, which are more accurate than those of other methods with contours. The European furniture IoU value is 0.534, which is higher than the generation effect of other methods and the CE value of noise floats up and down at 0.110. The 3D voxel reconstruction works best when the projection view angle K is 6, and the table works best at K=8 with 0.65. Overall, the model generated in this paper modeled 3D home products with high definition, good quality, and design.