首页|基于门控机制与重放策略的持续语义分割方法

基于门控机制与重放策略的持续语义分割方法

扫码查看
基于深度神经网络的语义分割模型在增量更新知识时由于新旧任务参数之间的干扰加之背景漂移现象,会加剧灾难性遗忘.此外,数据常常由于隐私、安全等因素无法被存储导致模型失效.为此,该文提出基于门控机制与重放策略的持续语义分割方法.首先,在不存储旧数据的情况下,通过生成对抗网络生成及网页抓取作为数据来源,使用标签评估模块解决无监督问题、背景自绘模块解决背景漂移问题;接着,使用重放策略缓解灾难性遗忘;最后,将门控变量作为一种正则化手段增加模型稀疏性,研究了门控变量与持续学习重放策略结合的特殊情况.在Pascal VOC2012数据集上的评估结果表明,在复杂场景10-2,生成对抗网络(GAN)、Web的设置中,该文在全部增量步骤结束后的旧任务性能比基线分别提升了3.8%,3.7%,在场景10-1中,相比于基线分别提升了2.7%,1.3%.
A Continual Semantic Segmentation Method Based on Gating Mechanism and Replay Strategy
Due to the interference and background drift between new and old task parameters,semantic segmentation model based on deep neural networks promotes catastrophic forgetting of old knowledge.Furthermore,information frequently cannot be stored owing to privacy concerns,security concerns,and other issues,which leads to model failure.Therefore,a continual semantic segmentation method based on gating mechanism and replay strategy is proposed.First,without storing old data,generative adversarial network and webpage crawling are used as data sources,the label evaluation module is used to solve the unsupervised problem and the background self-drawing module is used to solve background drift problem.Then,catastrophic forgetting is mitigated by replay strategy;Finally,gated variables are used as a regularization means to increase the sparsity of the module and study the special case of gated variables combined with continual learning replay strategy.Our evaluation results on the Pascal VOC2012 dataset show that in the settings of complex scenario 10-2,Generative Adversarial Networks(GAN)and Web,the performance of the old task after all incremental steps are improved by 3.8%and 3.7%compared with the baseline,and in scenario 10-1,they are improved by 2.7%and 1.3%compared with the baseline,respectively.

Continual learningSemantic segmentationReplay strategiesGating variables

杨静、何瑶、李斌、李少波、胡建军、溥江

展开 >

贵州大学机械工程学院 贵阳 550025

贵州大学公共大数据国家重点实验室 贵阳 550025

美国南卡罗莱纳州大学计算机科学与工程系 哥伦比亚29208

贵州民族大学机械电子工程学院 贵阳 550025

展开 >

持续学习 语义分割 重放策略 门控变量

国家自然科学基金贵阳市科技人才培养对象及培养项目国家重点研发计划Developing Objects and Projects of Scientific and Technological Talents in Guiyang City

62166005筑科合同[2023]48-8号2018AAA010800ZKH[2023]48-8

2024

电子与信息学报
中国科学院电子学研究所 国家自然科学基金委员会信息科学部

电子与信息学报

CSTPCD北大核心
影响因子:1.302
ISSN:1009-5896
年,卷(期):2024.46(7)