Pinhe Liu

About

Pinhe Liu is a researcher at the forefront of embedded artificial intelligence and robotic vision systems. His primary research areas encompass hardware-accelerated deep learning, binary neural networks, and real-time robotic perception. Liu’s major contribution lies in developing computationally efficient neural network architectures optimized for resource-constrained platforms. His most cited work introduces a novel Binary Fully Convolutional Neural Network (B-FCN) that leverages Taguchi method sub-optimization for precise floor region segmentation in complex indoor environments. This methodology is specifically designed for robotic vision, enabling accurate environmental understanding while dramatically reducing computational overhead. By implementing this network on SoC FPGA platforms, Liu demonstrates how hardware-software co-design can achieve real-time performance suitable for autonomous navigation. His approach addresses critical challenges in deploying deep learning on embedded systems, balancing accuracy with power efficiency. Though his citation count is still growing, Liu’s work represents an important step toward practical, low-power robotic vision systems that can operate effectively in dynamic, real-world settings.

Research Focus

Key Achievements

Top Papers

1. 1
   SoC FPGA Accelerated Sub-Optimized Binary Fully Convolutional Neural Network for Robotic Floor Region Segmentation

   5 citations · 2020

Key Collaborators

Contact & Links