• Technical Name
  • Indoor object detectiontrajectory prediction
  • Operator
  • Feng Chia University
  • Booth
  • Online display only
  • Contact
  • 謝瑞民
  • Email
  • dmshien@gmail.com
Technical Description We apply the deep learning algorithm to the camera in order to implement the real-time object detectionpedestrian tracking in the GPU development board. In addition, it is of the essence to maintain the accuracy ratethe lightweight model can be successfully executed on the PYNQ-Z2 to achieve real-time computing as well.
Scientific Breakthrough Use deep learning to detect indoor objectstrack predictions. Objects include pedestriansautomobiles. Use compression techniques to train the model to reduce the amount of parameters. We launched the Agile Model. Compared with Tiny-Yolo, the Model Size is reduced by 97.4, the execution speed is increased by 15FPS,the embedded platform TX2 reaches 30FPS,the AP can reach 93.5.
Industrial Applicability We provide the state-of-the-art technique of object detection, trajectory prediction,distance information. It can be applied to anti-collision warning functionincrease the convenience of automatic farm equipment, dronesetc. This application can reach our equipment maintenance demandsenhance the quality of public safety issues, such as airports, smart home devices,security