Daming Shi

ORCID: 0000-0003-4330-3338
Publications
Citations
Views
---
Saved
---
About
Contact & Profiles
Research Areas
  • Neural Networks and Applications
  • Image Processing and 3D Reconstruction
  • Handwritten Text Recognition Techniques
  • Image Retrieval and Classification Techniques
  • Advanced Computational Techniques and Applications
  • Face and Expression Recognition
  • Metallurgy and Material Forming
  • Advanced Neural Network Applications
  • Environmental and Agricultural Sciences
  • Video Analysis and Summarization
  • Cryospheric studies and observations
  • Optical measurement and interference techniques
  • Biometric Identification and Security
  • Face recognition and analysis
  • Advanced Measurement and Metrology Techniques
  • Climate change and permafrost
  • Multimodal Machine Learning Applications
  • Human Pose and Action Recognition

Huazhong University of Science and Technology
2025

Shenzhen University
2021

Kyungpook National University
2009

This study focuses on examining the characteristics of climate factors and extreme events in Northeast China during 1961–2010 by using daily data from 104 stations, including surface air temperature, precipitation, wind speed, sunshine duration, snow depth. Results show that annual mean temperature increased at a significant rate 0.35°C per decade, most notably Lesser Khingan Mountains winter. Annual rainfall had no obvious linear trend, while rainy days decreasing trend. So, rain intensity...

10.3724/sp.j.1248.2013.092 article EN cc-by-nc-nd Advances in Climate Change Research 2013-06-01

Semantic segmentation plays a key role in applications such as autonomous driving and medical image. Although existing real-time semantic models achieve commendable balance between accuracy speed, their multi-path blocks still affect overall speed. To address this issue, study proposes Reparameterizable Dual-Resolution Network (RDRNet) dedicated to segmentation. Specifically, RDRNet employs two-branch architecture, utilizing during training reparameterizing them into single-path inference,...

10.48550/arxiv.2406.12496 preprint EN arXiv (Cornell University) 2024-06-18

Off-line recognition of handwritten Chinese characters is considerable practical importance as well representing a very hard pattern problem. A popular approach to decompose into their component or ‘primitive’ parts – most usually strokes. Here, however, we take the less usual decomposing radicals. Active shape modelling applied and developed active radical modelling. In training, 60 examples each are represented by ‘landmark’ points, labelled semi-automatically, with radicals in different...

10.1076/jqul.10.1.41.16549 article EN Journal of Quantitative Linguistics 2003-04-01
Coming Soon ...