With the rise of automation and machine learning in our world, we have to work with a mindset to use it for our safety and for others' safety. This research paper aims to utilize two different machine learning models- K-Nearest Neighbors and Neural Networks - to try to find what is the best for vehicles to detect any oncoming dogs on the road. Through building several different models and comparing their overall accuracy, this research paper will answer the questions about the differences between the two models, and which one would be the most reliable for vehicles. The accuracy of the models are measured through several tests taking in 32x32 px images of both roads and dogs, and training and testing the different models.