+51 924-965-957

Deep Learning

  • Visión por Computación
  • Lima. Perú
  • Octubre 2019

Real-time object detection

Deep learning is trained so that learning is automatic, no rules are defined that must be followed; the training of the objects to be recognized is carried out from images that contain them. The images can come from a video or a video camera.
For training it is necessary to have a database of images that contains the objects we want to detect; The training process can last days and weeks, for this, computers with GPUs are used to significantly accelerate the training.
Object detector works from an IP camera, connected through our SDK for IP cameras that sends a frame that is processed by the trained model, returning the object to be detected. Programmed in c ++, with Qt and MFC compatible with windows 10.