A Research on Real-time Processing of High-Resolution Images with a small size FPGA

• Project/Area Number: 18K11209
• Research Category: Grant-in-Aid for Scientific Research (C)
• Allocation Type: Multi-year Fund
• Section: 一般
• Review Section: Basic Section 60040:Computer system-related
• Research Institution: University of Tsukuba
• Principal Investigator: MARUYAMA Tsutomu 筑波大学, システム情報系, 教授 (00292532)
• Project Period (FY): 2018-04-01 – 2021-03-31
• Project Status: Completed (Fiscal Year 2020)
• Budget Amount: ¥4,290,000 (Direct Cost: ¥3,300,000、Indirect Cost: ¥990,000); Fiscal Year 2020: ¥1,430,000 (Direct Cost: ¥1,100,000、Indirect Cost: ¥330,000); Fiscal Year 2019: ¥1,300,000 (Direct Cost: ¥1,000,000、Indirect Cost: ¥300,000); Fiscal Year 2018: ¥1,560,000 (Direct Cost: ¥1,200,000、Indirect Cost: ¥360,000)
• Keywords: FPGA / 画像処理 / 実時間処理

Outline of Final Research Achievements

We have implemented the NL-means algorithm on a small FPGA by using our new zigzag scan method. By this scan method, the computation time becomes double, but the required memory size for keeping the intermediate calculation result can be reduced to 1/50, and it becomes possible to realize real-time noise reduction using NL-means algorithm on a small size, namely, reasonably priced FPGA. We presented this result on an international conference (Parallel Computing 2019). We also researched how to reduce the programming complexity of this zigzag scan by using high level synthesis tools.

Academic Significance and Societal Importance of the Research Achievements

FPGAを用いた画像処理の高速化の研究は国内外で数多く行われている。しかし、これらの研究における主眼点は、より高速かつ高精度な処理の実現にあり、必要とされる回路量はそれ程重視されていない。本研究は、実時間処理の要求を満たした上で、高精度を維持しつつ回路量の削減を図るものであり、このような研究はほとんど行われていない。本研究で用いる手法の有効性は Box Filter に基づく手法に限られるが、Box Filter を用いたプログラムは様々な分野で数多く開発されており、本研究の成果を広い範囲に適用することが期待される。

Research Products

• [Presentation] An Implementation of Non-Local Means Algorithm on FPGA (2019)
  • Author(s): Hayato Koizumi, Tsutomu Maruyama
  • Organizer: Parallel Computing 2019
  • Int'l Joint Research