Tracking System Improved!

Our master student Jakob Mischek just finished his work on our tracking system which just got way better. Check out the video and his thesis (in German). The new system was made possible by an improved upstream  process, the detection and decoding of the bee markers got better thanks to work done by Leon Sixt. See this post for more information!

New Image Processing Pipeline Finished!

Our team member Leon Sixt has developed a new image processing component for the decoding of bee markers. The method is based on convolutional neural networks, the current state of the art in computer vision. ConvNets are awesome, however, they need large amounts of labeled data. In our case, this would have meant to click not only the image location of a tag but also the bit configuration for many thousand tag instances. Leons new solution “RenderGAN” (paper in prep) was to train a special network that can learn to reproduce realistic tag images. He then trained a decoder network using only generated tag images  – no human labeling necessary!

Tag Team Action!

We have marked all bees in our observation hive and the first full-fledged recording season has begun. We are very excited – everything is running smoothly so far (Knock on wood!!!).