Principles and Applications of Deep Convolutional Neural Networks

报告题目：Principles and Applications of Deep Convolutional Neural Networks

报告时间：2025 年3月21日下午 18:00

报告地点：南湖校区教学科研楼104室

主办单位：数学与统计学院

报告人：初雅莉

报告人简介：初雅莉，长春工业大学数学与统计学院 博士，研究方向为大数据分析

摘要：Deep Convolutional neural networks (DCNNs) are an important model in the field of deep learning and are widely used in computer vision tasks. DCNNs is composed of convolutional layer, pooling layer and fully connected layer. The convolutional layer realizes image feature extraction by means of local receptive field and weight sharing mechanism. The principle of feature extraction is based on hierarchical abstraction mechanism. In shallow layer of network, convolution kernel captures low-level features. With the deepening of the network level, the semantic object representation is gradually formed by nonlinear combination of low-level features. The pooling layer reduces the data dimension by downsampling. The fully connected layer integrates the feature vectors output by the pooling layer to complete the mapping from low-dimensional features to higher-level semantic space. Deep convolutional neural networks show significant advantages in the application fields of image migration and object recognition.