A CRF approach to fitting a generalized hand skeleton model

Abstract

We present a new point distribution model capable of modeling joint subluxation (shifting) in rheumatoid arthritis (RA) patients and an approach to fitting this model to posteroanterior view hand radiographs. We formulate this shape fitting problem as inference in a conditional random field. This model combines potential functions that focus on specific anatomical structures and a learned shape prior. We evaluate our approach on two datasets: one containing relatively healthy hands and one containing hands of rheumatoid arthritis patients. We provide an empirical analysis of the relative value of different potential functions. We also show how to use the fitted hand skeleton to initialize a process for automatically estimating bone contours, which is a challenging, but important, problem in RA disease progression assessment.

Serious Games Analytics

“…We developed a method to automatically label the bones in hand radiographs (Mihail, Blomquist, & Jacobs, 2014 ). The labeled set of bones (hereby referred to as a point distribution model, or more generally, shape) contains the positions and orientations of the long bones in a hand, and is robust to capture specifi c deformities caused by RA, namely joint space narrowing and subluxation.…”

“…We consider each radiograph a sample in RA hand shape space, after the transformation using the algorithm outlined by Mihail et al ( 2014 ). Hand shape is represented as a vector s (set of { x , y } coordinates, in some order, that specify where each bone is located on a radiograph.…”