基于深度学习的可回收垃圾分类算法

Recyclable Garbage Classification Algorithm Based on Deep Learning

基于深度学习可回收垃圾的分类算法是实现垃圾智能分类的关键。文章基于Image Net数据集的Inception模型,选用参数绑定与参数共享、提前终止、数据增强等多种方式防止数据过拟合,并在主干网络加入了注意力机制,提高了分类准确率。

The classification algorithm based on deep learning recyclable garbage is the key to realize intelligent garbage classification.This paper,based on the Inception model of Image Net data set,selects parameter binding and parameter sharing,early termination,data enhancement and other ways to prevent data overfitting,and adds attention mechanism to the backbone network to improve the classification accuracy.