不改变原意使其更有逻辑改写：The results of deep learning-based diabetic retinal classification have been remarkable in recent years Qiong Li et al 6 improved the AlexNet model by performing another batch normalizat

In recent years, deep learning-based diabetic retinal classification has shown remarkable results. Several studies have improved existing models and achieved high accuracy in detecting diabetic retinopathy. For instance, Qiong Li et al [6] improved the AlexNet model by adding a batch normalization layer and using migration learning strategy, achieving an accuracy of 93%. Kermany [7] used the migration learning algorithm for OCT image diagnosis, while Li et al. [8] developed a deep learning algorithm for DR detection based on color fundus images groups with high accuracy. De Fauw et al. [9] proposed a two-stage deep neural network model for OCT image diagnosis, achieving an accuracy of 95%. Hassan et al [10] proposed a deep convolutional neural network and a structural tensor based segmentation framework (CNN-STSF) for automatically segmenting eight retinal layers from normal and diseased OCT images. However, there are still some shortcomings, such as the use of overly complex backbones that may lead to low inference time or ignoring semantic information. To address these issues, we made further improvements to the existing backbone ResNest by enhancing its multi-scale representation ability and implementing strong feature expression capability with fewer parameters and calculation costs