China Algorithm Engineer at Beijing, China
Jing Hao is an accomplished algorithm engineer at Baidu Inc. with a Master’s degree in Mechanical Engineering from HuaZhong University of Science and Technology and a Bachelor’s degree from China University of Mining and Technology. Specializing in advanced AI algorithms for visual and image applications, he has designed UAV inspection algorithms for the State Grid and improved defect detection systems. Jing’s academic contributions include pioneering glass surface segmentation methods and publishing in IEEE TCSVT. Passionate about integrating AI in healthcare, he stays updated with the latest technologies and shares insights through his WeChat account, OAOA, with over 130 original posts. High Energy Physics
Professional Profiles
Education
Master’s Degree in Mechanical Engineering – HuaZhong University of Science and Technology Postgraduate Recommendation: September 2020 – June 2022 Weighted Average Score: 89.43 / 100 Rank: 66 / 205 Thesis Title: Research on Visual Assisted System of Engineering Vehicles based on Panoramic Imaging Bachelor’s Degree in Mechanical Engineering – China University of Mining and Technology Graduated with Honors: September 2016 – June 2020 Weighted Average Score: 87.49 / 100 Rank: 22 / 325 Thesis Title: Design of Light Bridge Crane and Remote Consol. High Energy Physics
Work Experience
Baidu Inc., Beijing, China Algorithm Engineer, Visual Video and Image Application Group (July 2022 – Present) Designed UAV inspection algorithms for metal defects in transmission electricity lines of the State Grid and factory safety inspection algorithms. Developed a training scheme based on CAE (Context Autoencoder) self-supervised pre-training algorithm for UAV inspection algorithms. Conducted special data preprocessing for defect data and applied the CAE pretraining algorithm upon the Vision Transformer (ViT) architecture. Released the “Power Vision Large Model Test and Verification Report” with the State Grid Intelligent Research Institute, achieving a 4.25% increase in the AP50 metric. Explored the application of LLM in factory safety inspection, achieving higher accuracy rates using the MiniGPT4 image-text multimodal foundation large model. High Energy Physics
Rewards
National Scholarship (2018) First-class Scholarship of HUST (2021) First-class Scholarship of CUMT (2019) Second-class Scholarship of CUMT (2017) Outstanding Student of CUMT (2018) Excellent Student Leader of Jiangsu Province (2019) High Energy Physics
Publications
- T-mamba: Frequency-enhanced gated long-range dependency for tooth 3d cbct segmentation, Publication date: 2024.
- Language-aware multiple datasets detection pretraining for DETRs, Publication date: 2024.
- GEM: Boost Simple Network for Glass Surface Segmentation via Segment Anything Model and Data Synthesis, Publication date: 2024.
- GEM: Boost Simple Network for Glass Surface Segmentation via Vision Foundation Models, Publication date: 2023.
- Simple parameter-free self-attention approximation, Publication date: 2023.
- A stronger stitching algorithm for fisheye images based on deblurring and registration, Publication date: 2023.
- A lightweight and accurate recognition framework for signs of X-ray weld images, Publication date: 2022.
- Heterogeneous Generative Knowledge Distillation with Masked Image Modeling
.