Abstract
Efficient architectures and implementations of median filters have been well investigated in the past. In this article, we focus on median filters for very big scientific applications with very large windows and an infinite stream of data, inspired by big data needs in the Square Kilometre Array (SKA) pulsar search engine, but transferable to other big data domains. We propose a novel approach for very large rectangular windows on an FPGA accelerator device able to support the processing of infinite streams of data. OpenCL is used for rapid parameter sweeping and design space exploration based on a pipelined model of the system. Evaluation on a host/accelerator system with an Arria 10 device surpassed 64 million values processed per second considered for the SKA real time requirement, achieving 83.4M value/s while reading from/writing to disk. These results are compared with a state-of-the-art software implementation only achieving 41M value/s for over twice the total system energy cost.
- [1] . 2012. A parallel edge preserving algorithm for salt and pepper image denoising. In Proceedings of the 2012 3rd International Conference on Image Processing Theory, Tools and Applications. 97–104.
DOI: DOI: Google ScholarCross Ref
- [2] . 1997. FPGA implementation of a median filter. In Proceedings of the IEEE Region 10 Annual Conference on Speech and Image Technologies for Computing and Telecommunications. Vol. 2, 437–440.
DOI: DOI: Google ScholarCross Ref
- [3] . 2008. Performance evaluation of image processing algorithms on the GPU. Journal of Structural Biology 164, 1 (2008), 153–160.
DOI: DOI: Google ScholarCross Ref
- [4] . 2009. Introduction to Algorithms. The MIT Press.Google Scholar
Digital Library
- [5] . 2012. OpenCL for FPGAs: Prototyping a compiler. In Proceedings of the International Conference on Engineering of Reconfigurable Systems and Algorithms. WorldComp, 1.Google Scholar
- [6] . 2013. SKA1 System Baseline Design.
Technical Report . SKA Office.Google Scholar - [7] . 2009. High-throughput one-dimensional median and weighted median filters on FPGA. IET Computers & Digital Techniques 3, 4 (2009), 384–394.
DOI: DOI: Google ScholarCross Ref
- [8] . 1993. Computing 2-D min, median, and max filters. IEEE Transactions on Pattern Analysis and Machine Intelligence 15, 5 (
May 1993), 504–507.DOI: DOI: Google ScholarDigital Library
- [9] . 2019. Design of a 2D median filter with a high throughput FPGA implementation. In Proceedings of the 2019 IEEE 62nd International Midwest Symposium on Circuits and Systems. 1073–1076.
DOI: DOI: Google ScholarCross Ref
- [10] . 1979. A fast two-dimensional median filtering algorithm. IEEE Transactions on Acoustics, Speech, and Signal Processing 27, 1 (
Feb. 1979), 13–18.DOI: DOI: Google ScholarCross Ref
- [11] . 2017. Pulsar searches with the SKA. Proceedings of the International Astronomical Union 13, S337 (2017), 171–174.
DOI: Google ScholarCross Ref
- [12] . 2010. Post-correlation radio frequency interference classification methods. Monthly Notices of the Royal Astronomical Society 405, 1 (2010), 155–167.
DOI: DOI: Google ScholarCross Ref
- [13] . 2021. Design and implement of median filter toward remote sensing images based on FPGA. In Proceedings of the 2021 IEEE 14th International Conference on ASIC. 1–4.
DOI: DOI: Google ScholarCross Ref
- [14] . 2007. Median filtering in constant time. IEEE Transactions on Image Processing 16, 9 (
Sept. 2007), 2389–2394.DOI: DOI: Google ScholarDigital Library
- [15] . 2014. Fine-tuned high-speed implementation of a GPU-based median filter. Journal of Signal Processing Systems 75, 3 (
1 June 2014), 185–190.DOI: DOI: Google ScholarDigital Library
- [16] . 1975. Applications of a nonlinear smoothing algorithm to speech processing. IEEE Transactions on Acoustics, Speech, and Signal Processing 23, 6 (
Dec. 1975), 552–557.DOI: DOI: Google ScholarCross Ref
- [17] . 2021. Scalable FPGA median filtering via a directional median cascade. In Proceedings of the 2021 IEEE 29th Annual International Symposium on Field-Programmable Custom Computing Machines. 273–273.
DOI: DOI: Google ScholarCross Ref
- [18] . 2010. A fast weighted median algorithm based on quickselect. In Proceedings of the 2010 IEEE International Conference on Image Processing. 105–108.
DOI: DOI: Google ScholarCross Ref
- [19] . 2013. Energy-efficient median filter on FPGA. In Proceedings of the 2013 International Conference on Reconfigurable Computing and FPGAs. 1–8.
DOI: DOI: Google ScholarCross Ref
- [20] . 2018. Median filtering with very large windows: SKA algorithms for FPGAs. In Proceedings of the 28th International Conference on Field Programmable Logic & Applications. IEEE.
DOI: DOI: Google ScholarCross Ref
- [21] . 2019. Frequently Asked Questions about the SKA. Retrieved 8 June, 2022 from https://www.skatelescope.org/frequently-asked-questions/.Google Scholar
- [22] . 2010. OpenCL: A parallel programming standard for heterogeneous computing systems. Computing in Science Engineering 12, 3 (
May 2010), 66–73.DOI: DOI: Google ScholarDigital Library
- [23] . 2018. Hierarchical histogram-based median filter for GPUs. Acta Polytechnica Hungarica 15, 2 (2018), 49–68.Google Scholar
- [24] . 2007. Efficient automatic noise reduction of electron tomographic reconstructions based on iterative median filtering. Journal of Structural Biology 158, 2 (2007), 196–204.Google Scholar
Cross Ref
- [25] . 2016. High-level designs of complex FIR filters on FPGAs for the SKA. In Proceedings of the 2016 IEEE 18th International Conference on High Performance Computing and Communications. 797–804.
DOI: DOI: Google ScholarCross Ref
- [26] . 2006. Fast median and bilateral filtering. In Proceedings of the ACM SIGGRAPH 2006 Papers. ACM, New York, NY, 519–526.
DOI: DOI: Google ScholarDigital Library
Index Terms
Median Filters on FPGAs for Infinite Data and Large, Rectangular Windows
Recommendations
A polyphase filter for GPUs and multi-core processors
Astro-HPC '12: Proceedings of the 2012 workshop on High-Performance Computing for Astronomy DateSoftware radio telescopes are a new development in radio astronomy. Rather than using expensive dishes, they form distributed sensor networks of tens of thousands of simple receivers. Signals are processed in software instead of custom-built hardware, ...
Multivariate Median Filters and Partial Differential Equations
Multivariate median filters have been proposed as generalizations of the well-established median filter for gray-value images to multichannel images. As multivariate median, most of the recent approaches use the $$L^1$$L1 median, i.e., the minimizer of ...
Performance of several types of median filters in spectral domain
Intelligent information processing IIMedian filter is well known for removing impulsive noise and preserving edges. Repeatedly filtering of any one-dimensional signal with a median filter will produce a root signal. Any impulses in the input signal will be removed by sufficient number of ...






Comments