Overall, we were happy with the system that we developed. It works quite well for its simplicity and lack of any more complicated techniques, though it is not super-simple. In the end, the features that we chose to collect ended up being good enough to identify between people. Indeed, the cepstrum is a very useful regime in which to analyze and characterize speech, and we feel it should be discussed at least once in ELEC 301 (even if only for five minutes). Likewise, after the great advice of [4], pitch information also proved itself to be useful. So in fact our project was a success.

Continuations
There are a number of ways that this project could be extended. Perhaps one of the most common tools in speaker identification is the Hidden Markov Model (HMM). This uses theory from statistics in order to (sort of) arrange our feature vectors into a Markov matrix (chains) that stores probabilities of state transitions. That is, if each of our codewords were to represent some state, the HMM would follow the sequence of state changes and build a model that includes the probabilities of each state progressing to another state. This is very useful, since it gives us another way of looking at speaker-unique characteristics.

There are also other weighting algorithms (and indeed VQ techniques) that we could have implemented, one of which is suggested in [5]. It would be an interesting study to determine the effectiveness of more complicated methods such as this one.