基于自适应图像增强和CNN的少数民族服饰图像分类研究

Research on Minority Clothing Classification Based on Adaptive Image Enhancement and Convolutional Neural Network

针对传统的少数民族服饰图像分类采用人工处理,无法满足信息化时代对图像自动分类的要求问题,提出了一种基于自适应图像增强和卷积神经网络的少数民族服饰图像分类算法。该算法通过计算复杂环境下少数民族服饰图像的平均亮度,采用不同的算法进行图像增强,提高图像对比度,降低噪声,并利用卷积神经网络学习不同少数民族服饰图像的特征,实现少数民族服饰图像的自动分类。实验结果表明,所提出的算法提高了准确率、召回率和F1值,能够有效地对白族、苗族、蒙古族、维吾尔族和藏族服饰进行分类和识别,为少数民族服饰的信息处理奠定了良好的基础,也为民族文化的传承和保护提供了支持。

To address the problem that the traditional image classification of ethnic minority costume images using manual processing cannot meet the requirements of automatic image classification in the information age,a classification algorithm of ethnic minority costume images based on adaptive image enhancement and convolutional neural network is proposed.The algorithm calculates the mean brightness of minority dress images in complex environment,uses different algorithms for image enhancement to improve image contrast and reduce noise,and uses convolutional neural networks to learn the features of different minority dress images to achieve automatic classification of minority dress images.The experimental results show that the proposed algorithm improves the accuracy,recall and F1 value,and can effectively classify and recognize the Bai,Miao,Mongolian,Uyghur and Tibetan costumes,which is a good foundation for the information processing of minority costumes and also supports the inheritance and protection of ethnic cultures.