Exploring fashion trends in color through enhanced MobileNetV3 model
Given the limitations of current methods for analyzing and predicting fashion trends in col-or based on data published by authoritative institutions,such as limited data sets,insufficient accura-cy,and poor real-time data,this paper proposes a research method for fashion color trends based on an improved MobileNetV3 model.The enhanced MobileNetV3 model is employed for rapid clothing classification tasks.Sales data from a time-sequenced e-commerce platform is used as samples,and the main colors of the clothing are analyzed based on the GrabCut algorithm.The proportions of the main colors and mixed colors are calculated using the K-means algorithm.The trend data for fashion trends in color is obtained by analyzing the main colors of clothing from three dimensions:time,clothing type,and brand.The study found that compared to traditional methods,the proposed research meth-od based on the enhanced MobileNetV3 model offers greater real-time data accuracy,increased capaci-ty and quicker analysis speed.
MobileNetV3GrabCutK-meanspopular colorprimary color extraction
NETL
NSTL
万方数据
