Browse by author
Lookup NU author(s): Chris Nash, Dr Rajesh NairORCiD, Dr Mohsen Naqvi
Full text for this publication is not currently held within this repository. Alternative links are provided below where available.
Attention Deficit Hyperactivity Disorder (ADHD) is one of the most prevalent mental health conditions in the world. If undiagnosed during childhood, there is a high chance of the individual leading a subpar livelihood. With low diagnosis rates and long wait times for a diagnosis appointment, there is motivation to support clinicians. This paper proposes a novel system that can analyse facial behaviour with machine learning to detect ADHD in humans. To analyse facial landmarks and movement of the facial features, a Long Short Term Memory (LSTM) network is exploited. We present a novel multi-modal dataset that contains 8 ADHD subjects and 12 controls with over 20 hours of recordings. By using this novel dataset, we identify facial behaviour differences between ADHD subjects and controls. By applying the facemesh on the video data, the facial behaviour difference between ADHD subjects and controls is also proposed. The proposed approach provides a validated average classification accuracy of 88.24%±3.93% for detecting ADHD. With diagnosis being a subjective decision from a clinician, we aim to introduce a robust objective measure with the proposed work.
Author(s): Nash C, Nair R, Naqvi SM
Publication type: Conference Proceedings (inc. Abstract)
Publication status: Published
Conference Name: International Conference on Artificial Intelligence Robotics, Signal and Image Processing (AIRoSIP)
Year of Conference: 2023
Online publication date: 10/08/2023
Acceptance date: 02/08/2023
URL: https://airosip.umy.ac.id/2023/
Notes: Received "Best Paper" Award
