人工智能时代的目标识别课程教学

Research on Teaching Target Recognition in the Era of Artificial Intelligence

目标识别的核心任务是判别图像的目标类型,通常采用机器学习算法完成,而鲁棒特征提取是其关键。传统人工提取的特征通常存在表达能力较弱的问题。深度学习理论的兴起为高精度目标识别提供了有效的解决途径,同时因其理论较为丰富,涉及面较广,给目标识别课程的讲授带来了新的挑战。通过设计具有专业特色的课程内容、构建知识链条、师生共建活跃课堂和全方位考核模式的探索等进行课程改革;在深度神经网络目标识别内容设计方面,侧重讲授关键概念、经典网络、研究进展及未来发展方向,为从事相关工作的教师和学生提供参考。

The core task of target recognition is to distinguish the type of target of an image.Target recognition is usually completed using machine learning algorithms,and robust feature extraction is its key.However,traditional manually extracted features often have weak expressive power.The rise of deep learning theory provides an effective solution for high-precision target recognition,and due to its rich theory and wide coverage,it also brings new challenges to the teaching of target recognition courses.Through designing the professional course content,constructing the knowledge chain,collaborative construction of active classrooms between teachers and students,and exploring comprehensive assessment models,teaching reform is carried out.In terms of explaining the content of deep neural network for target recognition,the focus is on introducing key concepts,classical networks,research progress,and future development directions,which may provide reference for teachers and students engaged in related work.