Automated bird phenotyping

For NIOO-KNAW, we developed an automated bird phenotyping application. This particular research project is on ‘changes in bill size in Parusmajor (great tits) in response to human provided food sources’. It will help in gaining a better understanding on how evolution works.

During this research, a large amount of pictures of birds are taken. These pictures are used to measure birds’ bill sizes. This is very time-consuming manual work. Therefore, we developed an application that automatically determines the bird’s bill size, and supports the researchers in managing their image processing workflow.

In this case, traditional computer vision techniques were used. But in further development, deep learning will be applied to also measure the curve of the bill. Also see our Automated seed germination measurements project, in which deep learning is already being used.