Research On The Soft Decoration Layout Method Based On Deep Learning

In recent years,people's living conditions have been significantly improved,and their living standards have also continued to pursue high quality.The home improvement industry has become a new type of consumer fashion.The architectural decoration(including design)industry in the domestic market is developing rapidly,and people's demand for home improvement design is also changing day by day.At the same time,the problems of both developments are also emerging.The standardization response of furnishing layout design is low,resulting in a lot of repetitive labor.The designer's design cost is expensive for ordinary people and the design cycle is long.In this paper,combined with actual needs,a furnishing layout method based on deep learning is proposed.This article applies deep learning to the layout design of furnishing,focusing on standardized furnishing layout design.Considering the large scale of the features of furnishing layout design,the features such as room structure,furnishing information and design style are extracted.Select the bedroom and living room as the main subject of the study.The basic process,characteristics and design concepts of the layout design of furnishing are analyzed in detail.The feature design of the furnishing design data set is targeted to increase the upper limit of the model and algorithm.Based on the BP neural network(back propagation),a soft layout model is constructed.The network structure,training process,and model parameter optimization of the model are analyzed.Batch Normalization,Dropout mechanism and other methods are used to speed up the training speed and convergence process,improve network stability,and prevent the model from overfitting.Analyze the experimental results by evaluating indicators.Aiming at the problem that the BP neural network has a slow learning rate and is prone to obtain local minimums,a deep belief network(Deep Belief Network)is used to establish a furnishing layout model.Combining the contrast divergence algorithm and the conjugate gradient method reduces the time required for network model reconstruction and improves prediction accuracy.Finally,the results of the experiment are analyzed,and the model is explored in the application of the actual furnishing layout.The experimental results show that the DBN-based furnishing layout model has better stability and more accurate prediction of the furnishing layout model than the BP neural network furnishing layout model.The deep learning-based furnishing layout is in practice It has good effect in application.