Scholarlink

<p>While recent deep learning-based object detection has achieved great success in various fields, it remains challenging to find tiny objects in aerial imagery on-the-fly using mobile devices. Since mobile platforms such as drones operate with limited onboard computing power, handling high-resolution images to find tiny objects with compute-intensive deep learning-based applications often fails to meet their real-time constraints. To mitigate this problem, we propose HashEye, a novel framework that enables fast on-drone tiny object detection by efficiently suppressing spatial redundancy in aerial imagery. HashEye utilizes a lightweight hashing algorithm to rapidly scan image patches; patches exhibiting high hash collision frequencies are identified as background and suppressed. Subsequently, the remaining salient patches are dynamically rearranged into a hardware-friendly dense image for efficient inference. Experimental results on two real-world datasets demonstrate that HashEye achieves up to a 5.25<InlineEquation ID="IEq1"><EquationSource Format="TEX">\(\times\)</EquationSource></InlineEquation> speedup compared to the baseline, maintaining detection capability.</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

HashEye: a real-time on-drone high-resolution tiny object detection via spatial pruning

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年5月6日

查看全文
  • Hyeonji Hong,
  • Nakyeong Lee,
  • Kwangwoo Jang,
  • Gague Kim,
  • Yangjae Jeong,
  • Chanyoung Oh

摘要

While recent deep learning-based object detection has achieved great success in various fields, it remains challenging to find tiny objects in aerial imagery on-the-fly using mobile devices. Since mobile platforms such as drones operate with limited onboard computing power, handling high-resolution images to find tiny objects with compute-intensive deep learning-based applications often fails to meet their real-time constraints. To mitigate this problem, we propose HashEye, a novel framework that enables fast on-drone tiny object detection by efficiently suppressing spatial redundancy in aerial imagery. HashEye utilizes a lightweight hashing algorithm to rapidly scan image patches; patches exhibiting high hash collision frequencies are identified as background and suppressed. Subsequently, the remaining salient patches are dynamically rearranged into a hardware-friendly dense image for efficient inference. Experimental results on two real-world datasets demonstrate that HashEye achieves up to a 5.25\(\times\) speedup compared to the baseline, maintaining detection capability.