It is undeniable that deep learning has opened up a plethora of opportunities for solving different problems in various fields. Within certain markets, in particular the video and entertainment industry, ML-based object detection can help increase accuracy and efficiency. However, it has also led to an emergence of a mindset that ML can solve many complex problems without much human intervention. The aim of this paper is to make it clear that there are limitations that prevent that from being the case, at least for the time being. A good understanding of the inherent limitations of the ML methods helps us devise ways to use the technology more effectively.
Read more →