Baoguang Shi
Principal Researcher at Microsoft
Redmond, Washington, United States
Summary
🤩
Rockstar🎓
Top SchoolBaoguang Shi is a Principal Researcher at Microsoft in Redmond with 11 years of experience specializing in computer vision and pattern recognition. He holds a PhD in Computer Vision from Huazhong University of Science and Technology and has published extensively (see Google Scholar), blending deep academic rigor with applied research. At Microsoft he leads vision research while maintaining hands-on contributions to influential open-source OCR and text-recognition projects such as CRNN and ASTER, fixing core CTC bugs and shaping attention-based decoders. His background includes a visiting stint at Cornell Tech under Serge Belongie and an internship at Microsoft Research Asia, reflecting strong academic-industry collaboration. Known for tackling practical model and data-preparation issues, he bridges research prototypes and robust implementations used in real-world image text recognition.
11 years of coding experience
Doctor of Philosophy (PhD) Computer Vision Pattern Recognition, Doctor of Philosophy (PhD) Computer Vision Pattern Recognition at Huazhong University of Science and Technology
English, Chinese
Github Skills (15)
attention-mechanism10
mask-rcnn10
model-building10
computer-vision10
faster-rcnn10
machine-learning10
lstm10
rnn-model10
tensorflow10
python10
n10
lua9
pytorch9
cprogramming-language7
c-language7
Programming languages (4)
Github contributions (5)
bgshih/aster
Dec 2017 - Mar 2021
Recognizing cropped text in natural images.
Role in this project:
Back-end Developer & ML Engineer
Back-end Developer & ML EngineerContributions:1 review, 121 commits, 2 PRs in 3 years 4 months
Contributions summary:Baoguang appears to be actively involved in developing the core architecture for text recognition in natural images, specifically focusing on attention-based models. The commits demonstrate work on defining and implementing an attention-based decoder, including modifications to the core decoder structure. The user also contributed to integrating a feature extractor and building the structure for a multi-predictor model, indicative of machine learning model development and integration.
deep-learningrecognitioncomputer-visionocrnatural
bgshih/CRNN
Dec 2015 - Dec 2015
Role in this project:ML Engineer
ML EngineerContributions:1 branch in 1 day
Contributions summary:Baoguang primarily contributed to bug fixes and updates within the CRNN (Convolutional Recurrent Neural Network) project. Their work involved modifications to the `create_dataset.py` script, the `ctc.cpp` file, and the `LstmLayer.lua` file. These changes likely addressed compatibility issues and improved the model's functionality. The user also updated `src/utilities.lua` to handle model state.
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.