Automatic Outcome in Manual Dexterity Assessment Using Colour Segmentation and Nearest Neighbour Classifier.

Area of Science:

• Rehabilitation Medicine
• Biomedical Engineering
• Movement Science

Background:

• Objective assessment of motor function is crucial for evaluating rehabilitation effectiveness.
• Traditional clinical scales like the Box and Blocks Test (BBT) assess manual dexterity.
• Previous work introduced an automated version (ABBT) using Microsoft Kinect.

Purpose of the Study:

• To evaluate the feasibility of the Automated Box and Blocks Test (ABBT) for automated manual dexterity assessment.
• To develop an improved algorithm for reliable automatic cube counting in the ABBT.
• To assess hand motor function in Parkinson's disease (PD) patients using the ABBT.

Main Methods:

• Developed an algorithm using CIELab color space image segmentation and Nearest Neighbor (NN) rule for cube counting.
• Conducted a pilot study on Parkinson's disease patients, performing three functional assessments.
• Compared manual BBT scores with automated ABBT counts and analyzed additional kinematic data.

Main Results:

• High correlation observed between manual BBT and automated ABBT scoring.
• Achieved an average success rate of 92% for automatic cube counting.
• ABBT provided additional data on cube displacement, including average velocity and detection time instants.

Conclusions:

• The ABBT is a feasible and reliable tool for automating the assessment of unilateral gross manual dexterity.
• Additional kinematic data from ABBT can serve as indicators of health status, coordination, and dexterity.
• ABBT offers valuable supplementary information beyond traditional scoring for user performance analysis.