Rapid pedobarographic image registration based on contour curvature and optimization

Abstract

Image registration, the process of optimally aligning homologous structures in multiple images, has recently been demonstrated to support automated pixel-level analysis of pedobarographic images and, subsequently, to extract unique and biomechanically relevant information from plantar pressure data. Recent registration methods have focused on robustness, with slow but globally powerful algorithms. In this paper, we present an alternative registration approach that affords both speed and accuracy, with the goal of making pedobarographic image registration more practical for near-real-time laboratory and clinical applications. The current algorithm first extracts centroid-based curvature trajectories from pressure image contours, and then optimally matches these curvature profiles using optimization based on dynamic programming. Special cases of disconnected images (that occur in high-arched subjects, for example) are dealt with by introducing an artificial spatially linear bridge between adjacent image clusters. Two registration algorithms were developed: a 'geometric' algorithm, which exclusively matched geometry, and a 'hybrid' algorithm, which performed subsequent pseudo-optimization. After testing the two algorithms on 30 control image pairs considered in a previous study, we found that, when compared with previously published results, the hybrid algorithm improved overlap ratio ( 010 . 0 = p ), but both current algorithms had slightly higher mean-squared error, assumedly because they did not consider pixel intensity. Nonetheless, both algorithms greatly improved the computational efficiency ( 8 25 ± ms and 9 53 ± ms per image pair for geometric and hybrid registrations, respectively). These results imply that registration-based pixellevel pressure image analyses can, eventually, be implemented for practical clinical purposes.

Registration of plantar pressure images

“…From the experimental results [10][11][12], we can point out the following conclusions. All algorithms revealed high accuracy, speed and noise robustness, being the most accurate the one presented in [12] that is based on a two-steps registration scheme and iterative optimization.…”

“…The registration algorithm presented in [10] is based on the matching of the external contours of the feet represented in the input images, and can be described in the following steps ( Figure 1): 1) extraction of the external contours; 2) computation of an affinity matrix between the contours' points; 3) searching for the best matching by minimizing the sum of the affinities; 4) computation of the geometric transformation that best register the matched points; 5) registration of the input images considering the transformation computed. …”

“…In that framework, for the pre-registration the user can choose the matching of the represented contours [10], the correlation of the images' phases or the cross-correlation of the input images [11]. Additionally, the Powell's method is adopted in the optimization process of the second step, taking into account one of the following (dis)similarity measures: the mean squared error (MSE), the mutual information (MI) or one based on the exclusive-or (XOR) [8].…”

“…In addition, pedobarographic image registration supports pixel-level statistics, which makes possible the extraction of biomechanically-relevant information more effectively than traditional regional techniques [3]. Several studies on the registration of pedobarographic images have been developed; for example, using principal axes transformation [4], modal matching [5,6], principal axes combined with steepest descent gradient search [7], optimization with evolutionary algorithms [8], based on foot size and progression angle [9], contours matching [10], optimization of the cross-correlation (CC) computed in the frequency domain [11], phase correlation [11], and optimization of an image (dis)similarity measure using an iterative scheme [12]. In this work, the later four methodologies are studied; thus, their fundamentals are introduced and a discussion about their results is presented.…”

Spatio-temporal alignment of pedobarographic image sequences

“…There are some studies on the alignment of pedobarographic image pairs; for example, those based on: principal axes transformation [6]; modal matching [3,17,23,24]; principal axes combined with a search based on the steepest descent gradient optimization algorithm [15]; optimization based on genetic algorithms [16]; foot size and foot progression angle [8]; matching the contours represented in the input images [13]; optimization of the cross-correlation or phase correlation computed in the frequency domain [11]; and using a hybrid approach that combines a feature based solution with an intensity based solution [12].…”