Book for learning statistics and probability theory to help developer deeper understanding of machine learning algorithms

I am new to Machine learning and I find it hard to understand many of the probability and statistical approaches. I don't have a proper probability and statistical background and I want to gain that understanding. Especially topics like markov chains, markov random fields [used a lot in computer vision] go over the top of my head. Can someone please recommend me some good books [assuming I have very little mathematical background] to cover these topics?