RECOGNITION AND INTELLIGENT DESIGN OF CHAIR SHAPE FEATURE LABEL UNDER DEEP LEARNING
In order to mine the effective information of big data images,accurately identify chair modeling features in massive photos and quickly design them according to feature tags,an intelligent recognition method of chair modeling features based on deep learning is proposed.Firstly,through literature research,it is determined that the labels for the chair modeling feature elements are selected based on typology and morphological analysis.Secondly,MoblieNet network is used to build a recognition and classification model of big data chair images,and the large number of chair images obtained are identified and screened through the image recognition model,and the characters are selected.Meet the required chair image;Finally,the multi-label learning model is used to establish an intelligent recognition and labeling model for chair shapes,verify the effectiveness of the model,and identify the labeled chair images.Finally,the chair image with a certain label can be selected to build a training dataset,and a new image can be generated by generating an adversarial network,which can be used as a design sketch to activate designer inspiration and better assist intelligent design.
万方数据
维普
