A customized CNN architecture for detection and classification of livestock medicine vials

Abstract

Computer vision is a widely-used technique for many applications such as quality control, surveillance, autonomous vehicles as well as in the medical field. It is very versatile and can be applied, in this case, to help automate the process of keeping track of the medicines being administered to livestock. Automating this process is important for various reasons. The first is to reduce human error when creating a written record of the medicines being administered to the livestock, it takes away the responsibility of recording this information from the people administering it so they can focus solely on administering the medicines. Secondly, it could also help a livestock farm keep track of their medicine inventory as well as alerting them of medicines that are close to their expiry date in order for them to buy new ones. Finally, in combination with the previous point, the system could be used on various farms to keep track of the inventory at each one and automatically purchase medicines when inventory runs low or medicines are near their expiry date.

9