Haruyoshi Yonekawa

Papers

Total Citations

159

H-Index

About

Haruyoshi Yonekawa is a researcher specializing in embedded computer vision, deep learning acceleration, and real-time object detection systems. His work sits at the intersection of hardware design and machine learning, with a particular focus on making neural network-based detection algorithms efficient enough for deployment on resource-constrained platforms such as robotics, autonomous vehicles, security cameras, and drones. Yonekawa's most recognized contribution is his 2018 paper "A Lightweight YOLOv2," which has accumulated 135 citations and addresses the dual challenge of spatially localizing objects through bounding box regression while simultaneously classifying them — all within real-time frame rates. This work has proven especially influential in the embedded systems community, where computational efficiency is paramount. His earlier 2017 paper on multiscale sliding window object detection using a fully pipelined binarized convolutional neural network on an FPGA (24 citations) demonstrates his pioneering efforts in hardware-software co-design, leveraging binarized networks to achieve high performance under strict power and area constraints. Together, these works reflect Yonekawa's enduring commitment to bridging the gap between state-of-the-art deep learning and practical, deployable embedded vision systems — a contribution of growing relevance in today's edge computing landscape.