微型电脑应用2025，Vol.41Issue(4)：50-54,5.
基于深度学习的鸟类细粒度图像分类方法
Bird Fine-grained Image Classification Method Based on Deep Learning
摘要
Abstract
Due to the lack of image noise processing in the process of bird fine-grained image classification,the accuracy of image classification is low and the classification time is long.Therefore,a bird fine-grained image classification method based on deep learning is proposed.The fine-grained image of birds is preprocessed by gray normalization to eliminate the noise in the image.The fine-grained image features of birds after preprocessing are extracted by entropy to obtain the detailed information of image texture thickness and layout structure.The extracted fine-grained image features of birds are taken as data samples,and sent to the sample layer of probabilistic neural network for training.The image classifier is designed by the channel attention mecha-nism to classify fine-grained images of birds.The experimental test shows that the bird fine-grained image classification method based on deep learning has good effect and high classification efficiency.关键词
概率神经网络/鸟类细粒度图像/灰度归一化/特征提取Key words
probabilistic neural network/bird fine-grained image/gray normalization/feature extraction分类
信息技术与安全科学引用本文复制引用
王代远,宋伟奇,孔丽云..基于深度学习的鸟类细粒度图像分类方法[J].微型电脑应用,2025,41(4):50-54,5.基金项目
2021年教育部科技发展中心重点项目(2020ITA03005) （2020ITA03005）
广西自然科学基金重点项目(2017GXNSFDA198028,2018GXNSFDA138006) （2017GXNSFDA198028,2018GXNSFDA138006）
