关键词 · 相关资讯与研究

#WGAN-GP

沿着“WGAN-GP”继续探索,共 0 条内容。

暂未找到相关内容

可以换一个关键词,或按农业领域继续浏览。

浏览农业主题

你可能还想看

没有这个标签,内容接近
论文Computers and Electronics in Agriculture一手来源

A knowledge-informed WGAN-GP framework for farm-scale soil organic matter mapping under sample scarcity

A knowledge-informed WGAN-GP framework for farm-scale soil organic matter mapping under sample scarcity。Computers and Electronics in Agriculture

关注点

该框架将领域知识嵌入WGAN-GP生成器,以缓解农场尺度土壤有机质制图的样本稀缺问题;其适用性取决于小样本条件下生成数据的可信度与迁移范围。

阅读 0
论文Smart Agricultural Technology一手来源

Self-Supervised Learning With RGB-Guided Attention and Individual Consistency for Pig Weight Estimation

Self-Supervised Learning With RGB-Guided Attention and Individual Consistency for Pig Weight Estimation。Smart Agricultural Technology

关注点

该方法用RGB引导注意力并结合个体一致性约束做自监督猪只体重估测,可减少对人工称重标注的依赖;适用边界在于需验证不同猪舍光照与拍摄角度下的泛化性。

阅读 1
论文Computers and Electronics in Agriculture Vol.250 111938 2026-08

Addressing representation instability, strategy rigidity, and class imbalance in active learning for agricultural pest and disease recognition(解决农业病虫害识别主动学习中表征不稳定、策略僵化和类别不平衡问题)

为降低农业病虫害识别深度学习部署的专家标注成本,文章提出三层主动学习框架:动态原型记忆通过动量指数滑动平均保持稳定类别表征;上下文感知样本选择采用神经上下文老虎机动态调整分支组合比例;类别平衡评分通过逆频加权在选择阶段直接优先关注少数类。在IP06(害虫)和Natural CW-10(杂草)数据集上,框架在90%数据可用率下分别取得83.36%和85.07%的准确率,在20%低数据可用率下较最强基线分别领先4.12和2.68个百分点。结合GAN增强数据后,IP06上准确率提升至84.91%,Natural CW-10上提升至86.42%,证明主动学习与生成式AI在智能农业中的互补价值。

阅读 10
论文Computers and Electronics in Agriculture一手来源

AutoGHRW: A knowledge-based interpretable automated machine learning framework for grasshopper risk warning

AutoGHRW: A knowledge-based interpretable automated machine learning framework for grasshopper risk warning。Computers and Electronics in Agriculture

关注点

该框架将领域知识与自动化机器学习结合用于蝗虫风险预警,并强调可解释性;适用性需结合具体蝗区数据验证,目前仅为期刊论文层面的方法报告。

阅读 0