Scholarlink

This paper examines the overall performance of two system mastering models, Time series help Vector Machines (SVMs) and Recurrent Neural Networks (RNNs), on classification of Hyper Spectral photographs (HSIs). The fashions are in comparison based totally on their recognition accuracy which includes the time requirements to build fashions. The paper additionally evaluates two varieties of preprocessing techniques: minimum Redundancy maximum Relevancy (MRMR) and precept aspect evaluation (PCA). The effects of the experiments highlight that SVMs and RNNs perform considerably distinct for HSI class and that counting on one version simplest may lead to exclusive type performance. Moreover, the experiments show that with increasing complexity, the quantity of time required to assemble the RNNs will increase significantly. The comparisons in the paper allow further exploration in the discipline and help the choice of the high-quality model with the aid of capacity practitioners and developers.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Comparing Time Series Assist Vector Machines and Recurrent Neural Networks for Hyper Spectral Image Popularity

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Shubhashish Goswami,
  • Afroz Pasha,
  • Awakash Mishra,
  • Ramkumar Krishnamoorthy

摘要

This paper examines the overall performance of two system mastering models, Time series help Vector Machines (SVMs) and Recurrent Neural Networks (RNNs), on classification of Hyper Spectral photographs (HSIs). The fashions are in comparison based totally on their recognition accuracy which includes the time requirements to build fashions. The paper additionally evaluates two varieties of preprocessing techniques: minimum Redundancy maximum Relevancy (MRMR) and precept aspect evaluation (PCA). The effects of the experiments highlight that SVMs and RNNs perform considerably distinct for HSI class and that counting on one version simplest may lead to exclusive type performance. Moreover, the experiments show that with increasing complexity, the quantity of time required to assemble the RNNs will increase significantly. The comparisons in the paper allow further exploration in the discipline and help the choice of the high-quality model with the aid of capacity practitioners and developers.