Examining the Integration of Digital Technology for Ankle-Foot Orthosis Registration and Quantitative Shape Comparison
Abstract
Ankle-foot deformities can affect an individual’s gait and cause increased pain and discomfort during the activities of daily living. This can significantly affect quality of life and lead to further complications and health-related risks. Custom-made ankle-foot orthotics (AFOs) are widely used in patient care and rehabilitation to improve and restore ankle-foot positioning, support, mobility, and stability. The traditional approach in designing custom-made AFOs relies heavily on the experience and craftsmanship of the orthotist during shape capture, design, and fabrication [1].
Digital workflows in orthotics, including design and fabrication of AFOs, are advancing patient care and rehabilitation by improving the quality and consistency of assistive devices [2]. These workflows include a series of steps that automate tasks and processes within an organization. Recent breakthroughs in digital technology have the potential of transforming the design processes of orthoses by replacing traditional methods. Digital workflows typically involve using an optical scanner to capture the foot shape, and computer-aided design and computer-aided manufacturing (CAD/CAM) systems to optimize the shape and fabricate the AFO. The use of digital technology requires less physical labor, allows models to be easily stored, modified, and replaced, and increases design flexibility. However, it remains a challenge to translate the subjective skills of orthotists into the digital realm [3]. Creating a visual representation of the areas of shape change during manual fabrication helps quantify the design process. This is often referred to as digital mapping. Registration is the process of digitally overlaying or superimposing two shapes with respect to each other. This is necessary to compare, or digitally map, different types of AFO workflows and designs. Current registration methods require the use of anatomical landmarks or reflective markers as reference points for aligning two shapes [4]. This process is inefficient and tedious. Therefore, this study serves to investigate and validate a suitable registration method that will streamline the process of aligning digital AFO models for shape comparison. This proposed digital algorithm will be used to align and compare AFO models acquired using both traditional and digital shape capture methods.
The constraints and limitations of current registration methods were identified by systematically testing the alignment digital models. An optimized registration algorithm using both iterative closest points (ICP) and principal component analysis (PCA) was developed and evaluated to align digital AFO models. Evaluations were conducted by aligning digital AFO models with known deformations to their original shapes. While ICP and PCA were ineffective on their own, results showed that these methods could be used together to obtain accurate registration for shapes misaligned in both translation and rotation.
In collaboration with the Department of Orthotics and Prosthetics (O&P) at the Holland Bloorview Kids Rehabilitation Hospital (HBKRH), participants prescribed a unilateral or bilateral AFO will be recruited. The participant will undergo traditional and digital shape capture in the same session. Scanning the ankle and foot directly will allow for alignment corrections to be applied digitally before quantitatively comparing the shapes. This work is essential in adopting full digital workflows for AFOs. It will also help improve digital technologies for orthotic and prosthetic practice.
