深度学习中弱监督细粒度识别方法与应用综述
首发时间:2019-10-16
摘要:本文归纳介深度学习下基于弱监督的细粒度识别方法与应用综述深度学习下基于弱监督的细粒度识别方法与应用综述绍深度学习时代弱监督下的细粒度识别众多算法,从弱监督网络结构、弱监督定位、特征融合方法与损失函数改进与优化等方面进行了具体的论述,分析了相关算法的优缺点和应用场景,最终结合目前比较流行的神经网络结构搜索技术和视频识别技术讨论了细粒度识别的学术前沿研究点和应用前景。
For information in English, please click here
A Survey of Weakly Supervised Fine-Grained Recognition Methods and Applications in Deep Learning
Abstract:This paper introduces the many algorithms of fine-grained recognition under weak supervision in the deep learning era, and discusses the weak supervision network structure, weak supervision positioning, feature fusion method and loss function improvement and optimization, and analyzes the advantages and disadvantages of related algorithms. And the application scenario, finally combined with the current popular neural network structure search technology and video recognition technology to discuss the academic frontier research points and application prospects of fine-grained recognition.
引用
No.****
动态公开评议
共计0人参与
勘误表
深度学习中弱监督细粒度识别方法与应用综述
评论
全部评论0/1000