An intelligent system for image-based rating of corrosion severity at stem taper of retrieved hip replacement implants. (November 2018)

• Record Type: Journal Article
• Authors: Milimonfared, Roohollah; Oskouei, Reza H.; Taylor, Mark; Solomon, Lucian B.
• Is Part Of: Medical engineering & physics. Volume 61(2018)
• Page Start: 13
• Page End: 24
• Publication Date: 2018-11
• Subjects: Total hip arthroplasty -- Metallic implants -- Machine learning -- Digital image processing -- Texture analysis; Biomedical engineering -- Periodicals; Biomedical Engineering -- Periodicals; Physics -- Periodicals; Génie biomédical -- Périodiques; Biomedical engineering; Electronic journals; Periodicals; 610.28
• DOI: 10.1016/j.medengphy.2018.08.002
• Languages: English
• ISSNs: 1350-4533

Abstract

Highlights: An intelligent algorithm was developed to rate corrosion damage on hip implants. Neighbourhood Component Analysis selected features with higher discriminatory power. Local texture features can be used to characterise images of corroded regions. Bayesian optimisation was used to fine-tune hyperparameters of SVM classifier. Fivefold cross-validation accuracy of 85% for rating corrosion damage on implants. Abstract: Visual scoring of damage at taper junctions is the sole method to quantify corrosion in large-scale retrieval studies of failed hip replacement implants. This study introduces an intelligent image analysis-based method that objectively rates corrosion at stem taper of retrieved hip implants according to the well-known Goldberg scoring method. A Support Vector Machine classifier was used that takes in vectors of global and local textural features and assigns scores to the corresponding images. Bayesian optimisation fine-tunes the hyperparameters of the classifier to minimise the cross-validation error.