Hi Grace! That's really cool- its amazing how innovation sparked by a simple idea can have such large influence! CCD is unfortunately still a significant problem, but scientists have made some headway in finding its causes so it has been out of media coverage for a while. My app measures risk percentage for CCD based on its possible causes, predicting a hive's likelihood of getting the disorder in the future. CCD in itself is not an illness- it is a phenomenon that arises from other diseases such as Varroa Mite Infestation. Thus, SafeHive has to identify symptoms of these factors in order to assess CCD risk.
Hi Adithi! Thanks for reaching out! It's great to see that you guys are working on the social aspect of this issue to increase general CCD awareness. Regarding your questions, my algorithm primarily focuses on classifying several types of illnesses visible in both the bees and their honeycombs, considering that these are currently the most indicative signs of the disorder. While pesticides and pathogens are thought to have a potential role in causing CCD, they are generally visually indistinguishable. This means that including them as features in the convolutional neural network will decrease its overall accuracy since it is unable to correctly identify their images. The alternative to using Computer Vision, receiving information directly from the beekeepers regarding pesticides in their area, would be largely unfeasible because most are accidentally transmitted across long distances. Therefore, it is more effective to look for bees affected by pathogens such as Israeli Acute Paralysis rather than for the pesticide or virus itself.