Accelerating the Fast Hadamard Single-Pixel Imaging (FHSI) on GPUs: Techniques and Optimizations
Author:
Quero Carlos Osorio1, Durini Daniel1, Rangel-Magdaleno Jose1, Ramos-Garcia Ruben1, Martinez-Carranza Jose1
Affiliation:
1. Instituto Nacional de Astrofisica Optica y
Electronica
Abstract
Abstract
The recent advancements in edge computing power are primarily attributable to technological innovations enabling accelerators with extensive hardware parallelism. One practical application is in computer imaging (CI), where GPU acceleration is pivotal, especially in reconstructing 2D images through techniques like Single-Pixel Imaging (SPI). In SPI, compressive sensing (CS) algorithms, deep learning, and Fourier transformation are essential for 2D image reconstruction. These algorithms derive substantial performance enhancements through parallelism, thereby reducing processing times. These techniques fully utilize the potential of the GPU by implementing several strategies. These include optimizing memory accessed, expanding loops for efficiency, designing effective computational kernels to reduce the number of operations, using asynchronous operations for better performance, and increasing the number of actively running threads and warps. In lab scenarios, integrating embedded GPUs becomes essential for algorithmic optimization on SoC-GPUs. This study focuses on quickly improving the Fast Hadamard Single-Pixel Imaging (FHSI) for 2D image reconstruction on Nvidia's Xavier platform. By implementing various parallel computing techniques in PyCUDA, we managed to speed up the process by approximately 10 times, significantly reducing processing times to nearly real-time levels.
Publisher
Research Square Platform LLC
Reference63 articles.
1. Osorio Quero, Carlos A. and Durini, Daniel and Rangel-Magdaleno, Jose and Martinez-Carranza, Jose (2021) {Single-pixel imaging: An overview of different methods to be used for 3D space reconstruction in harsh environments}. Review of Scientific Instruments 92(11): 111501 https://doi.org/10.1063/5.0050358, https://pubs.aip.org/aip/rsi/article-pdf/doi/10.1063/5.0050358/15597770/111501\_1\_online.pdf, https://doi.org/10.1063/5.0050358, 0034-6748, {Different imaging solutions have been proposed over the last few decades, aimed at three-dimensional (3D) space reconstruction and obstacle detection, either based on stereo-vision principles using active pixel sensors operating in the visible part of the spectra or based on active Near Infra-Red (NIR) illumination applying the time-of-flight principle, to mention just a few. If extremely low quantum efficiencies for NIR active illumination yielded by silicon-based detector solutions are considered together with the huge photon noise levels produced by the background illumination accompanied by Rayleigh scattering effects taking place in outdoor applications, the operating limitations of these systems under harsh weather conditions, especially if relatively low-power active illumination is used, are evident. If longer wavelengths for active illumination are applied to overcome these issues, indium gallium arsenide (InGaAs)-based photodetectors become the technology of choice, and for low-cost solutions, using a single InGaAs photodetector or an InGaAs line-sensor becomes a promising choice. In this case, the principles of Single-Pixel Imaging (SPI) and compressive sensing acquire a paramount importance. Thus, in this paper, we review and compare the different SPI developments reported. We cover a variety of SPI system architectures, modulation methods, pattern generation and reconstruction algorithms, embedded system approaches, and 2D/3D image reconstruction methods. In addition, we introduce a Near Infra-Red Single-Pixel Imaging (NIR-SPI) sensor aimed at detecting static and dynamic objects under outdoor conditions for unmanned aerial vehicle applications.}, 11 2. Sorna, Anumeena and Cheng, Xiaohe and D'Azevedo, Eduardo and Won, Kwai and Tomov, Stanimire (2018) Optimizing the Fast Fourier Transform Using Mixed Precision on Tensor Core Hardware. 10.1109/HiPCW.2018.8634417, 3-7, , , 2018 IEEE 25th International Conference on High Performance Computing Workshops (HiPCW) 3. Ma, Mengchao and Sun, Qianzhen and Gao, Xicheng and Wang, Guan and Deng, Huaxia and Zhang, Yi and Guan, Qingtian and Zhong, Xiang (2021) {High-efficiency single-pixel imaging using discrete Hartley transform}. AIP Advances 11(7): 075211 https://doi.org/10.1063/5.0057119, https://pubs.aip.org/aip/adv/article-pdf/doi/10.1063/5.0057119/12850514/075211\_1\_online.pdf, https://doi.org/10.1063/5.0057119, 2158-3226, {Single-pixel imaging technology is popular with invisible wavelengths and low light environments. However, the time-consuming steps hindered the development of single-pixel imaging technology. To improve imaging efficiency, a high-efficiency one-step single-pixel imaging method based on the discrete Hartley transform is proposed. The proposed method does not require a large number of fringe patterns and only requires a real-number calculation. The number of fringe patterns required for the proposed method is only half of that required for the four-step phase-shift Fourier method at the same sampling rate. Although a one-step method, it also uses the idea of differential measurements and adds upsampling processing strategies, which simultaneously improve the signal-to-noise ratio of the recovered image. The simulation shows that the peak signal-to-noise ratio and structural similarity index of the recovered target scene exceed 20 dB and 80\%, respectively, when the sampling rate is 30\%. Only 20 164 patterns are needed to reconstruct a (256 × 256)-pixel image. After defocusing the gray stripe pattern into a binary pattern, it only takes milliseconds to project these patterns into the target. It can be seen that the experimental results of the proposed method are significantly better than those of the two-step phase-shift method under dramatical noise interference. With the rapid development of advanced equipment, this method will represent significant progress in the real-time reconstruction of single-pixel imaging.}, 07 4. Bowman, Nathaniel and Carrier, Erin and Wolffe, Greg (2013) PyGASP: Python-based GPU-accelerated signal processing. 10.1109/EIT.2013.6632683, 1-6, , , IEEE International Conference on Electro-Information Technology , EIT 2013 5. Murthy, Giridhar Sreenivasa and Ravishankar, Mahesh and Baskaran, Muthu Manikandan and Sadayappan, P. (2010) Optimal loop unrolling for GPGPU programs. 10.1109/IPDPS.2010.5470423, 1-11, , , 2010 IEEE International Symposium on Parallel & Distributed Processing (IPDPS)
|
|