3rd year projects 2016 - 17

  • JPB-UG-2a: Acoustic Scene Classification (Mark Lister)
  • JPB-UG-2b: Acoustic Scene Classification (Danny Heard)
  • JPB-UG-3: Sound event detection in real and synthetic audio (Jack Deadman)
  • JPB-UG-4: Domestic audio tagging (Urh Krzic)
  • JPB-UG-8: Sonification of text data and clustering music patterns to match human language (Mikhail Molotkov)

Mail all

The project descriptions below are only intended as starting points. If you wish to discuss possibilities in greater detail I encourage you to email me to arrange a meeting.


JPB-UG-2a: Acoustic Scene Classification


JPB-UG-2b: Acoustic Scene Classification

Description

Acoustic scene classification is the task of taking an audio recording and assigning it to one of a number of predefined classes that characterise the environment in which it was recorded. For example, classes might be “cafe”, “street junction”, “office”. This project will build and evaluate novel solutions to this problem. We will adopt the evaluation framework provided by the DCASE2016 machine listening challenges.

Further reading

For further details see here,

Prerequisites

  • An interest in AI / machine listening
  • Python programming skills
[TOP]


JPB-UG-3: Sound event detection in real and synthetic audio

Description

Sound event detection is the task of trying to detect the presence of a certain sound (e.g. a door slamming) in a continuous audio recording. This problem is a key component of `scene understanding’. Solutions have many potential application, e.g. searching and retrieval of audio recordings, machine listening for autonomous robots, audio-based monitoring and surveillance systems. This project will build and evaluate novel solutions to this problem. We will adopt the evaluation framework provided by the DCASE2016 machine listening challenges.

Further reading

For further details see here,

Prerequisites

  • An interest in AI / machine listening
  • Python programming skills
[TOP]


JPB-UG-4: Domestic audio tagging

Description

Audio tagging is the process of attaching descriptive tags to short segments of audio. Typically the tags come from a small pre-defined set. This task is a key component of `scene understanding’. Solutions have many potential application, e.g. searching and retrieval of audio recordings, machine listening for autonomous robots, audio-based monitoring and surveillance systems. This project will build and evaluate novel solutions to this problem. We will adopt the evaluation framework provided by the DCASE2016 machine listening challenges which focuses on audio tagging for audio recorded in a family home.

Further reading

For further details see here,

Prerequisites

  • An interest in AI / machine listening
  • Python programming skills
[TOP]


JPB-UG-8: Sonification of text data and clustering music patterns to match human language

Description (student-propsed project)

Everyone is familiar with data visualisation tools. However, what if we were able to represent data in musical form?

This project aims to convert text data into music by using NLP techniques to extract specific patterns from the text – such as, what type of the words are used more in the text? how complex is the sentence structure? etc, – and then mapping those patterns to a specific set of music characteristics, e.g. high/low pitch, tempo, combination of the notes. The next step of the project is to construct a computer model that would used unsupervised machine learning techniques to automatically find similar patterns in human language and music and then to cluster this data in a way that each word would be mapped to a unique music combination. The tool would allow new ways of examining texts that might prove useful for plagiarism detection, authorship attribution, etc.

Current research in sonification area: http://www.economist.com/news/science-and-technology/21694992-scientific-data-might-be-filled-important-things-waiting-be-discovered. There is also a book that covers the main principles and ideas of sonification: “Auditory Display: Sonification, Audification and Auditory Interfaces” by Gregory Kramer.

[TOP]