Despite it not being possible to build the AI Blind Spot Detection, I still want to give an example of how it would be created. As I mentioned earlier in blog 23 (which can be read here), for AI to be able to recognise an input, it first has to be taught what one looks like.
For example, if you wanted to teach a machine to understand the difference between a cat and a dog, you’d first have to show it hundreds of different photos of each, training it to recognise what they look like. After a while, when shown either a cat or a dog, the computer will be able to identify which animal it is. This is the same approach I’d have to implement to get blind spot detection working. However, due to Coronavirus, I was never able to create the model. Instead, I ended up taking a slightly different approach to the technology.
Despite all this, I still want to give an example of the types of photographs that would be used to teach the machine. Down below are four pictures, each time taken from a blind spot position. Hundreds of similar pictures would be uploaded into the model, and after a while, the computer would be able to recognise any blind spot.
Commentaires