About Me

Personal Information:

• Name: Jingyu Yan
• Occupation: Deep Learning Programmer
• Email: tunmxy@163.com
• Location: China Xiamen
• Field of Expertise: Experienced in image-related technologies, including face recognition, license plate recognition, and object detection. Also skilled in audio processing, lightweight deployment solutions, and software programming.

Educational Background:

• Bachelor’s Degree:
  • University: Xiamen University of Technology
  • College: School of Software Engineering
  • Major: Software Engineering
  • Focus: Machine Learning, Software Design, Java Programming, Android Programming
• Master’s Degree:
  • University: Kharkiv National V.N. Karazin University
  • College: School of Mathematics and Computer Sciences
  • Major: Computer Sciences
  • Research Focus: Machine Learning, Programming Theory, Functional Programming, Stochastic Methods

Technical Skills:

• Programming Languages: Python, C/C++, Java, Scala
• Frameworks and Tools: TensorFlow, Pytorch, Numpy, GCC, CMake, MNN, RKNN, ONNX
• Professional Skills: Designing computer vision solutions, model training and tuning, mobile quantization, pruning, and deployment, as well as adapting and deploying on various embedded devices with NPUs.

Open-Source Projects:

• HyperLPR(5.4k Stars):
  • Link: https://github.com/szad670401/HyperLPR
  • Description: High-performance Chinese license plate recognition based on deep learning, capable of running on multiple platforms, including but not limited to Windows, Linux, MacOS, and various embedded adaptations.
• HyperLandmark(1.6k Stars):
  • Link: https://github.com/szad670401/HyperLandmark
  • Description: HyperLandmark is an efficient open-source facial keypoint tracking algorithm based on deep learning. It supports Android, iOS, and various embedded platforms, providing precise facial contour descriptions with 106 keypoints while maintaining outstanding accuracy in various lighting conditions. Notably, on the Qualcomm 820 (st) platform, it achieves a processing time of just 7 seconds for single-frame facial tracking.

Published Work:

• Deep Learning Techniques and Applications (2020, China Railway Publishing House), Xiaodong Zheng, Wei Zhu, Jingyu Yan, Weidong Xiao;
• A snoring detection method and apparatus based on multiple model stacking (CN202310231634.7), Wei Lu, Jingyu Yan, Liangtao Deng, Ying Huang, China, published;
• A snoring detection method and apparatus (CN202310231634.7), Liangtao Deng, Jingyu Yan, Ying Huang, Wei Lu, China, published.