Automatic Characterizations of Lumbar Multifidus Muscle and Intramuscular Fat with Fuzzy C-means based Quantization...

Area of Science:

• Medical Imaging
• Computer Vision
• Biomedical Engineering

Background:

• Low back pain (LBP) affects half the population, incurring high socioeconomic costs.
• Chronic LBP rehabilitation is crucial, with lumbar multifidi muscles and intramuscular fat being key areas of interest.
• Current ultrasonography analysis of these muscles is subjective due to human intervention.

Purpose of the Study:

• To develop an automated computer vision system for lumbar multifidi muscle thickness measurement.
• To analyze intramuscular fat distribution within the lumbar multifidi using an automated approach.
• To reduce operator subjectivity in analyzing lumbar multifidi muscles from ultrasound images.

Main Methods:

• Image processing algorithms enhance contrast for accurate muscle thickness measurement.
• Fuzzy C-Means (FCM) clustering is employed for intermuscular fat analysis.
• The system processes DICOM format ultrasound images.

Main Results:

• The automated system achieved promising results in computing lumbar multifidi thickness.
• 72% of cases showed minimal discrepancy (less than 0.2 cm) compared to human experts.
• FCM-based fat analysis demonstrated superiority over conventional histogram methods.

Conclusions:

• The proposed computer vision system offers an objective and accurate method for lumbar multifidi analysis.
• This automated approach can significantly aid in the clinical and research aspects of low back pain rehabilitation.
• The system provides a cost-effective and reliable tool for assessing muscle and fat characteristics relevant to LBP.