- Multimedia Data Analysis: We aim to develop methods that analyse a large amount of multimedia data (e.g., images, videos, music and text) to accurately identify relevant data to a certain meaning.
- Sensor-based Human Activity Recognition: We aim to develop methods that can accurately recognise human activities based on data, which are obtained from sensors embedded in wearable devices and environments.
- Multimodal Analysis: We aim to develop methods that analyse relations between different types of data, in order to improve or realise the recognition of semantic information that is difficult to recognise only with a single type of data.
- Human Perception-based Analysis: We aim to develop methods that extract meaningful information from a large amount of data by adopting human perception mechanisms, such as memory, adaptivity and planning. The developed methods can be utilised as fundamental technologies in the above-mentioned three topics.
Some past research topics can be found here