Afbeelding
Corrosie-project-

Few shot ship corrosion detection 

As part of the Professorship of Applied Sciences in Computer Vision and Data Science

Goal

Detect the corrosion of metal in ships. 

Main challenge  

The large variation in the visual appearance of corrosion combined
with the availability of only a few example images requires advanced few
shot models. 

Result 

A few-shot model that recognized ship corrosion from only a few training
images. 

Future

Apply the few-shot approach to other context in which only a limited amount
of image data is available.