Scholarlink

<p>Traditional filters based on full window consider that neighboring pixels surround the target pixel, which likely crosses over the potential edges. Although the side window (SW) with eight sub-windows alleviates artifacts and halos, it still faces a challenge with mandatory smoothing. To address the issue, a novel and simple steerable side window (SSW) framework is proposed in the paper. The proposed SSW utilizes the derivative of the Gaussian function to generate a set of side windows, including half- and quarter-windows with arbitrary orientation, effectively improving the perception of detailed information and significantly enhancing alignment with the pixel distribution. Meanwhile, the SSW can also be easily embedded in a variety of filters. It not only inherits the properties of SW, but also extends the flexibility of SW. Experiments are carried out on different filters, and the results show that the proposed SSW technique is both effective and efficient, enabling more advantageous applications.</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

The design and application of steerable side window framework

  • 期刊论文

  • 发表日期：

    2026年2月25日

查看全文
  • Xiaohong Jia,
  • Tao Lei,
  • Yingbo Wang,
  • Xuejun Zhang,
  • Guanghui Yan,
  • Asoke K. Nandi

摘要

Traditional filters based on full window consider that neighboring pixels surround the target pixel, which likely crosses over the potential edges. Although the side window (SW) with eight sub-windows alleviates artifacts and halos, it still faces a challenge with mandatory smoothing. To address the issue, a novel and simple steerable side window (SSW) framework is proposed in the paper. The proposed SSW utilizes the derivative of the Gaussian function to generate a set of side windows, including half- and quarter-windows with arbitrary orientation, effectively improving the perception of detailed information and significantly enhancing alignment with the pixel distribution. Meanwhile, the SSW can also be easily embedded in a variety of filters. It not only inherits the properties of SW, but also extends the flexibility of SW. Experiments are carried out on different filters, and the results show that the proposed SSW technique is both effective and efficient, enabling more advantageous applications.