Masayuki Shimoda

Papers

Total Citations

H-Index

About

Masayuki Shimoda is a researcher specializing in embedded systems, FPGA-based hardware acceleration, and deep learning implementation for real-time computer vision applications. His work sits at the critical intersection of neural network optimization and resource-constrained computing, addressing the practical challenge of deploying sophisticated AI models on compact, energy-efficient hardware. Shimoda's most notable contribution is his pioneering FPGA implementation of YOLOv2, the influential real-time object detection framework, demonstrating that high-accuracy, high-speed detection is achievable on embedded platforms — work that has garnered 18 citations and holds significant relevance for robotics, autonomous vehicles, and security systems. Building on this, he proposed a tri-state weight convolutional neural network approach tailored for FPGAs, offering improved performance-per-power efficiency without sacrificing detection quality. His 2020 research on fast monocular depth estimation further extends his contributions to 3D scene understanding on embedded devices, targeting applications in home robotics and drone navigation. With a cumulative citation impact spanning computer vision, neural network compression, and hardware design, Shimoda's research provides valuable frameworks for engineers and researchers seeking to bring deep learning capabilities to low-cost, real-world embedded environments.