基于生成对抗网络和视觉-语义对齐的零样本害虫识别方法  

Zero-Shot Pest Identification Based on Generative Adversarial Networks and Visual-Semantic Alignment

作　　者：李天俊 杨信廷[2] 陈晓 胡焕 周子洁 李文勇[2] LI Tianjun;YANG Xinting;CHEN Xiao;HU Huan;ZHOU Zijie;LI Wenyong(College of Information,Shanghai Ocean University,Shanghai 201306,China;National Engineering Research Center for Information Technology in Agriculture,Beijing 100097,China;College of Information Technology,Jilin Agriculture University,Jilin 130118,China)

机构地区：[1]上海海洋大学信息学院,上海201306 [2]北京市农林科学院信息技术研究中心,北京100097 [3]吉林农业大学信息技术学院,吉林长春130118

出　　处：《智慧农业（中英文）》2024年第2期72-84,共13页Smart Agriculture

基　　金：国家重点研发计划子课题(2022YFD2001801-2);北京市农林科学院院协同创新中心建设专项课题。

摘　　要：[目的/意义]害虫的精准识别对农作物虫害有效防治具有重大意义。然而,当前的害虫图像识别模型都是针对闭集数据构建的模型,难以对训练集中没有出现过的害虫种类(不可见害虫)进行推理,导致在实际应用过程中遇见不可见类别昆虫时误判现象尤为严重。[方法]针对这一问题,提出了一种适用零样本学习(Zero-Shot Learning,ZSL)和广义零样本学习(Generalized Zero-Shot Learning,GZSL)范式的害虫图像识别方法VSAWGAN,可以实现对可见(训练集中包含的类别)与不可见害虫种类的辨识。该方法基于生成对抗网络(Genera⁃tive Adversarial Network,GAN)生成伪视觉特征,将零样本害虫识别问题转化为传统监督学习任务,且通过引入对比学习来优化生成器的生成质量,添加视觉-语义对齐模块进一步约束生成器,使其能生成更具判别性的特征。[结果与讨论]在自建的一个适用于零样本学习研究的害虫数据集和几个公开数据集对提出方法进行了评估。其中,在多个公开数据上取得了目前最优结果,相比之前方法最大提升达到2.8%;在自建20类害虫数据集上取得了零样本设置下77.4%的识别精度和广义零样本设置下78.3%的调和精度,相比之前方法分别提升了2.1%和1.2%。[结论]所提方法能有效地将害虫的视觉特征泛化到不可见类,实现害虫的零样本识别,有助于提升害虫识别模型的泛化能力,为农作物新虫害的发现与防治提供帮助。[Objective]Accurate identification of insect pests is crucial for the effective prevention and control of crop infestations.However,existing pest identification methods primarily rely on traditional machine learning or deep learning techniques that are trained on seen classes.These methods falter when they encounter unseen pest species not included in the training set,due to the absence of image samples.An innovative method was proposed to address the zero-shot recognition challenge for pests.[Methods]The novel zero-shot learning(ZSL)method proposed in this study was capable of identifying unseen pest species.First,a comprehensive pest image dataset was assembled,sourced from field photography conducted around Beijing over several years,and from web crawling.The final dataset consisted of 2000 images across 20 classes of adult Lepidoptera insects,with 100 images per class.During data preprocessing,a semantic dataset was manually curated by defining attributes related to color,pattern,size,and shape for six parts:antennae,back,tail,legs,wings,and overall appearance.Each image was annotated to form a 65-dimensional attribute vector for each class,resulting in a 20×65 semantic attribute matrix with rows representing each class and columns representing attribute values.Subsequently,16 classes were designated as seen classes,and 4 as unseen classes.Next,a novel zero-shot pest recognition method was proposed,focusing on synthesizing high-quality pseudo-visual features aligned with semantic information using a generator.The wasserstein generative adversarial networks(WGAN)architecture was strategically employed as the fundamental network backbone.Conventional generative adversarial networks(GANs)have been known to suffer from training instabilities,mode collapse,and convergence issues,which can severely hinder their performance and applicability.The WGAN architecture addresses these inherent limitations through a principled reformulation of the objective function.In the proposed method,the contrastive module was design

关 键 词：害虫识别 语义知识 图像特征 生成对抗网络 对比学习 广义零样本学习 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术] S433[自动化与计算机技术—计算机科学与技术]