Image and Natural Language Processing module (CS41004)

Image and natural language processing is a way for computers to understand visual inputs and human languages.

Image processing concerns itself with the analysis and manipulation of digital images to enhance or extract information from the images. Computer vision (CV) has many real-world use cases, such as QR code recognition, software for self-driving vehicles, and tools for medical diagnosis.

Natural language processing (NLP) studies how computers can understand, interpret, and generate human languages. NLP is used in many state-of-the-art applications, including language translation, predictive text and semantic analysis, and AI chatbots.

As these technologies are increasingly integrated into society, the skillset required to develop CV and NLP applications is highly sought after. They allow organisations to improve their applications and processes by making them more user-friendly, efficient, and safe.

What you will learn

In this module, you will:

• study common computer vision (CV) tasks and methods
• investigate convolutional neural networks as a method for image processing
• study common natural language processing (NLP) tasks and methods
• investigate sequence modelling and recurrent networks as methods for NLP
• learn about transformer models for CV and NLP tasks
• explore image-to-text and text-to-image tasks
• examine multimodal processing, an AI paradigm that incorporates data from multiple input types such as images and text
• discuss how to evaluate the performance of methods for specific CV and NLP tasks

By the end of this module, you will be able to:

• demonstrate your understanding of ket tasks and challenges in CV and NLP
• explain modern methods used for CV and NLP, including those based on machine learning
• select appropriate methods to use for specific tasks in CV and NLP
• apply methods for CV and NLP using industry standard software frameworks
• evaluate and appraise the performance of methods for specific CV and NLP tasks

Assignments / assessment

• computer vision application development (20%)
• natural language processing application development (20%)
• written exam (60%)

Teaching methods / timetable

You will learn by taking a hands-on approach. This will involve taking part in tutorials and practical sessions.

Learning material is provided through videos, review notes, examples, and tutorial questions.