Kaiming Fu

Education

• Ph.D. in Electrical and Computer Engineering, University of California, Davis
• Double-Major M.S. in Statistic, University of California, Davis

Projects

• Annual Farm Robotics Challenge
  • Team Leader. Grand Prize Winner among National-wide Universities and Colleges
    • Designed a DepthAI-based real-time monitor system within a harvesting assistant robot, capable of autonomously tracking human operators and providing immediate posture analysis feedback, enhancing worker safety and efficiency through a custom Human Monitoring System.
    • Optimized agricultural productivity by enabling the robot to transport harvested crops directly to storage, effectively eliminating the reliance on manual tractor transportation.
• ”Inceptio-Tsinghua AIR Cup” Autonomous Driving Challenge
  • 1st Prize Winner among 1067 Teams
    • Utilized an Xbox controller to collect driving data for training a neural network with Imitative Learning, collaborating on semi-trailer acceleration control with the LCA lane keeping system.
    • Employed a range of advanced problem-solving techniques, including two-way search, greedy algorithms, space pruning, convex optimization, and the deep reinforcement learning PPO algorithm.
• Fine-Grained Classification in Plant Pathology
  • IEEE / CVF Computer Vision and Pattern Recognition Conference (CVPR) workshop
    • Preprocessed imbalanced dataset using a data sampler, employed ResNet50 as the baseline, achieving an F1 score of 0.70.
    • Implemented a Generative Data Augmentation model for image augmentation and training dataset balance.
    • Improved F1 score to 0.874 using a UNet-ResNet generator and a DenseNet discriminator.