Scholarlink

<p>PerceptPath is a new computational framework for intelligent journey route planning and display optimization. Integrating Dijkstra’s algorithm with a perceptual visual attention model helps the system to improve user engagement and immersion. This method generates Gaze Allocation Sequences via a Locality-maintained and Human-Like Active Learning module to prioritize semantically and visually salient path segments. At the same time, Dijkstra’s algorithm determines the shortest trip paths. These sequences are deeply aggregated using a deep aggregation network to extract high-level characteristics; subsequently, a Gaussian Mixture Model models them to allow adaptive visual retargeting. By means of dynamic modification of trip route displays depending on perceptual attention patterns, this technique preserves both functional relevance and aesthetic quality. Comparative analysis and user studies among empirical assessments show PerceptPath improves display precision by 3.66% and test time by 52.34% compared to traditional approaches. Combining computing efficiency and perceptual relevance greatly increases AI-driven route planning and immersive display optimization in hybrid reality settings.</p>

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

PerceptPath: travel path planning and display algorithm optimized by Dijkstra’s algorithm and a perceptual visual attention model

  • 期刊论文

  • 发表日期：

    2025年12月5日

查看全文
  • Dandan Lu,
  • Wei Lo,
  • Chun-Ming Yang

摘要

PerceptPath is a new computational framework for intelligent journey route planning and display optimization. Integrating Dijkstra’s algorithm with a perceptual visual attention model helps the system to improve user engagement and immersion. This method generates Gaze Allocation Sequences via a Locality-maintained and Human-Like Active Learning module to prioritize semantically and visually salient path segments. At the same time, Dijkstra’s algorithm determines the shortest trip paths. These sequences are deeply aggregated using a deep aggregation network to extract high-level characteristics; subsequently, a Gaussian Mixture Model models them to allow adaptive visual retargeting. By means of dynamic modification of trip route displays depending on perceptual attention patterns, this technique preserves both functional relevance and aesthetic quality. Comparative analysis and user studies among empirical assessments show PerceptPath improves display precision by 3.66% and test time by 52.34% compared to traditional approaches. Combining computing efficiency and perceptual relevance greatly increases AI-driven route planning and immersive display optimization in hybrid reality settings.