Semantic-based electronic recognition photo data amplification method and system

The invention relates to a semantic-based electronic recognition photo data amplification method and system, and the method achieves the considerable improvement of the number and semantic coverage through increasing the semantic coverage rate of a sample and the distribution uniformity of the sample for an amplified electronic recognition photo data set. And outputting the electronic recognition photo test data set meeting the preset quality requirement, thereby realizing amplification of a large number of high-quality data sets through a small number of data sets. According to the technical scheme provided by the invention, after the electronic recognition photo data is amplified, the semantic coverage rate and the sample distribution uniformity of the sample are also considered, so that the quality of the output data is better, various scene application requirements of deep learning can be met, the operation is simple, the accuracy rate is high, the user experience degree is good, and the satisfaction degr