Dongjiu Zhang

About

Dongjiu Zhang is a researcher at the forefront of computer vision and deep learning, with a primary focus on RGB-D image processing and 3D scene understanding. His most-cited work, "Fast Generation of High-Fidelity RGB-D Images by Deep Learning With Adaptive Convolution" (2020, 9 citations), introduces a novel deep-learning framework that transforms low-resolution, incomplete data from consumer-level RGB-D cameras into high-fidelity, high-resolution images. By developing adaptive convolution operators that intelligently handle missing regions and sparse inputs, Zhang’s method achieves both speed and accuracy—critical for real-time applications in robotics, augmented reality, and autonomous navigation. This contribution addresses a fundamental bottleneck in depth sensing, enabling more reliable 3D reconstruction from affordable hardware. Beyond this paper, Zhang’s research advances efficient neural architectures for multimodal data fusion, bridging the gap between raw sensor outputs and practical, high-quality visual outputs. His work is recognized for its practical impact, offering a scalable solution that democratizes access to high-fidelity depth imaging. For students and researchers, Zhang exemplifies how targeted algorithmic innovations can transform everyday hardware into powerful tools for perception and interaction.

Research Focus

Key Achievements

Top Papers

1. 1
   Fast Generation of High-Fidelity RGB-D Images by Deep Learning With Adaptive Convolution

   9 citations · 2020

Key Collaborators

Contact & Links