Scholarlink

<p>This work presents a comprehensive survey of Field-Programmable Gate Array (FPGA) implementations of image processing filters within Very Large-Scale Integration (VLSI) environments. The survey examines widely used filtering techniques, including bilateral, Gaussian, median, Finite Impulse Response (FIR), and adaptive filters, and compares them based on reported performance, power efficiency, and area utilization across different FPGA platforms. A fine-grained analysis of recent implementations indicates that FIR and Gaussian filters are commonly associated with lower reported execution times (as low as 2.34–7.31&#xa0;ns) and efficient resource utilization, whereas bilateral and adaptive filters, despite higher computational complexity, are widely reported to offer stronger edge preservation and noise suppression. The survey further highlights architectural trade-offs among FPGA families such as Spartan, Artix, Virtex, and Zynq, with emphasis on their suitability for real-time embedded imaging applications. Finally, emerging trends in AI-assisted FPGA design, dynamic reconfiguration, and energy-aware architectures are discussed as future research directions for advanced image processing systems.</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

Performance Analysis of FPGA-Based Digital Filter Architectures for Image Processing Applications

  • 期刊论文

  • 发表日期：

    2026年3月7日

查看全文
  • C. Saranya,
  • Kumar Ponnusamy,
  • R. Ramya,
  • R. Jagadeesan,
  • D. Poornakumar

摘要

This work presents a comprehensive survey of Field-Programmable Gate Array (FPGA) implementations of image processing filters within Very Large-Scale Integration (VLSI) environments. The survey examines widely used filtering techniques, including bilateral, Gaussian, median, Finite Impulse Response (FIR), and adaptive filters, and compares them based on reported performance, power efficiency, and area utilization across different FPGA platforms. A fine-grained analysis of recent implementations indicates that FIR and Gaussian filters are commonly associated with lower reported execution times (as low as 2.34–7.31 ns) and efficient resource utilization, whereas bilateral and adaptive filters, despite higher computational complexity, are widely reported to offer stronger edge preservation and noise suppression. The survey further highlights architectural trade-offs among FPGA families such as Spartan, Artix, Virtex, and Zynq, with emphasis on their suitability for real-time embedded imaging applications. Finally, emerging trends in AI-assisted FPGA design, dynamic reconfiguration, and energy-aware architectures are discussed as future research directions for advanced image processing systems.