NCTU

people detection base on deep learning

2017/09/13
CodeM17-CCL-10-1
Professorprofessor Cheng-Chang Lien
ApplicationCloud computing and the application of internet of things(IoTs), Surveillance and security, and Vehicle
FunctionImage detection and Image recognition
Technical BenefitDecrease error rate / improve stability, Expand monitoring scope, Improve system integration, Improve system intelligence, and Weather adaption
Technology StatusCan be transferred

Traditional pedestrian identification technology still cannot improve the recognition accuracy because of lighting variation, clothing color, complex background, human action and occlusion problems. Hence, it is very difficult to design a pedestrian identification system that can overcome the abovementioned problems. Therefore, we apply the deep learning technology to extract image features from local to global. The simulation results show the proposed DNN based pedestrian identification technology can recognize the pedestrians with the accuracy above 90%.

 (a)

(b)

(a) Original image form PIROPO database. Fig. (b) The result using deep learning.

Spec. request

  • platform: Windows 10
  • tools:Visual studio 2015
  • relative softwares: Opencv 2.4.9 、 cuda 7.5

 

demand

  • at least Intel Core 4
  • Ram: at least 4GB

VGA:Geforce GTX 1050Ti