Object detection is a concept that has been handed over to the machines for some time now. It all started with pattern recognition back in the 1960’s and the concept has been evolving ever since. Now it has come to such an extent that machines are successfully able to achieve real-time object detection. But all this isn’t done on its own and requires human help and instantiation in order to start working independently. Whenever there is a need for instantiation, there comes a need for an algorithm. Talking about algorithms, an algo is a finite series of steps which when performed lead to an outcome. Ever since the beginning of time, we have been using algorithms to do our daily chores. Ever since the beginning of machines, we have been using algorithms to program them and to define how they would react to various inputs. Similarly, when the concept of object detection came in, the series of development of algorithms began. There have been multiple developments over the past which have been able to do the job of object detection but the difference between these has always been of speed and accuracy as well as efficiency. Earlier algorithms were performed on still images that were captured out of cameras or taken out of video footages. Now, highly efficient algorithms have made it possible to provide real-time object detection that can be performed by continuously moving video footages.   Object detection is not a cakewalk and has a lot of challenges that it comes with. The goal to be achieved is overcome some challenges and reduce the effect of others to the bare minimum. It is true that there are trade-offs involved in this process but based on the objective of the program, these trade-offs can be easily managed. The main challenges that come up are mentioned below.