Skip to content

Vector object recognition based on YoloV3 neural network

Notifications You must be signed in to change notification settings

ROC5COR/Vector_AI_Vision

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

3 Commits
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Vector_AI_Vision

Bringing advanced computer vision on the Anki Vector robot. By using Vector's API to fetch image data and TinyYoloV3 neural network, vector is able to recognize objects in a scene.

Vector image

How does it works ?

  • Vector's API are used to fectch image data from Vector's camera
  • Images are preprocessed (reshaped,...)
  • Preprocessed images are sent to TinyYoloV3, which is a smaller version of Yolo optimized to reduced computational needs (which means higher framerates for less computational power)
  • Output of Yolo are decoded
  • Results are printed on Vector's screen + the captured frame from the camera
  • Results can also be spoken by Vector (which is cool in a first place but annoying finally)
  • And the process goes over and over again

Technical needs

References

About

Vector object recognition based on YoloV3 neural network

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages