University of Malta

A BCI for Rapid Image Searching
UOM Main Page
Apply - Admissions 2016
Campus Map button
CBS Facebook


A Brain-Computer Interface for Rapid Image Searching

Lead Investigator: Ms Elysia Calleja
Supervisor: Dr Owen Falzon, Centre for Biomedical Cybernetics
Co-Supervisor: Prof. Ing. Kenneth P. Camilleri,  Centre for Biomedical Cybernetics and Dept. of Systems and Control Engineering, Faculty of Engineering


Even though there have been great advances in computer vision systems, no system has come close to replicating the complexity of the human vision system for object detection. Humans can recognize objects of interest at a glance, even when the objects are shown under different angles and lighting. The recognition of a target object evokes an identifiable brain activity pattern in an individual, which can be recorded using electroencephalography (EEG). This pattern can be used to increase the efficiency of object detection by using the human vision system for object recognition, and computer processing power to analyse the EEG data and determine whether an object of interest was shown.

The aim of this project was to implement a brain-computer interface (BCI) to decode EEG data and determine objects of interest from a series of images shown at a high rate by using rapid serial visual presentation (RSVP). The system presents to the user a series of stimuli, consisting both target and non-target images, while EEG data is recorded. The recorded data was processed and different feature extraction methods were used to classify the data into target or non-target images. The different feature extraction methods analysed were the decimation method and three other methods in which a t-test is performed on the training data to find the maximally discriminable points from the EEG data window. The first method uses all the points from the t-test result (APT method); the second method uses only consecutive points from the result (CPT method); whilst the the last method uses the mean of the consecutive points (MCPT method). A linear discriminant analysis (LDA) classifier was used to classify the data into targets or non-targets.


No statistical difference was found for the target detection rate between the four methods used, however the decimation method performed the best with a statistical difference for the non-target detection rate. Nevertheless, the other methods proved to perform better with a smaller training data. With the available training data, the decimation method provided positive results with a target detection rate of 75% and non-target detection rate of 86%. The system provided positive results for both target detection and non-target detection rate, however the results can be greatly improved by using different classification techniques. Once the system provides sufficient results, a real-time system can be implemented. The real-time application with such a system can be used for various different applications in which the aim would be to reduce the object detection time as much as possible. 

Other Links

BCI research at the Dept. of Systems and Control Engineering [Link] 

Funding Award for BrainApp

CBC awarded research funding under the FUSION Technology Development Programme (TDP) 2017

New peer-reviewed journal article

CBC publishes in Biomedical Physics & Engineering Express

CBC attends EMBC 2017

Three Papers Accepted for Presentation at the Annual Conference of the IEEE Engineering in Medicine and Biology Society (EMBC 2017)

Research at UoM: Brain Computer Interface

Communicating using brain signals

Last Updated: 11 May 2017

Log In back to UoM Homepage