PE1-6.5 AI Applications in Signals, Images & Speech
You can pre-process and extract features from signal and time series data, apply neural networks to image classification, and describe speech production, sound classification and HMM-based speech recognition.
This is the unit that justifies the course sitting in an ECE department rather than a CS one: the inputs are signals, and feature extraction is signal processing by another name. The lesson worth carrying is that classical feature engineering and learned representations are two answers to one question, and which wins depends on how much data you have. Speech recognition via HMM closes the loop back to Unit 2 and shows a single model doing real work end to end.
Work through these
Pre-processing and feature extraction for signal and time series data
Preparing signals before any learning happens, which is where an engineer with domain knowledge adds the most value. Good features beat a better model on this kind of data.
Fundamentals of image processing for machine learning
The image operations a machine learning pipeline actually needs. It is a subset of a whole subject, chosen for what comes next.
Feature mapping
Turning raw input into a representation a model can use. It is the same idea as the item above, generalised.
Case study: image classification using artificial neural networks
The worked example that ties the network topic to real data. Doing it end to end is the point rather than the accuracy achieved.
Process of speech production: a block diagram approach
How speech is made, as a block diagram, which is the model every speech technique assumes. The source and filter split is the idea to carry.
Classification of speech sounds
The categories of speech sound and what distinguishes them acoustically. It is what makes the features in a recogniser meaningful rather than arbitrary.
Speech recognition using hidden Markov models
Recognition using the time-structured model from the knowledge topic. It closes the course by connecting two of its halves on a real application.
Sign in to keep your progress.
Free resources
We haven't checked most of these for screen reader use yet.
Links last checked 30 Aug 2026.
Stuck here?
Ask a mentor. A real person answers, and they can see exactly which topic you're on. Usually within a couple of working days.
Checking your session…
Topics shown in module order.