In-vivo evaluation of intervertebral kinematics can provide precious information for widespread spinal pathologies such as back pain, whiplash, that still lack of certain diagnoses. Analysis of fluoroscopic sequences screening spine tracts (e.g. lumbar, cervical) during unconstrained patient motion can be used to estimate vertebrae and segmental motion: even if limited, the 2D analysis can be employed to study motion onto sagittal plane. Estimation of vertebral kinematics relies on recognition of vertebrae position and rotation on each radiological frame; this can be achieved identifying specific feature points or landmarks. Manual selection results tedious and imprecise, automatic vertebrae recognition can be based on image template matching. This study proposes a particular template matching that uses smoothed image derivatives, which enhances main vertebral body outline. Vertebra location result more accurate and precise with respect to previous techniques. Results were tested against known data of a reference calibration model: the root mean square error resulted 0.2 degree for vertebral angles and 0.3 mm for vertebra positions. A further comparison was performed using previous findings obtained by processing real sagittal, lumbar fluoroscopic sequences: the root mean square error resulted 1.2 degree for vertebral angles and 0.8 mm for vertebra positions

Vertebrae tracking through fluoroscopic sequence: a novel approach

BIFULCO, PAOLO;CESARELLI, MARIO;ROMANO, MARIA;CERCIELLO, TOMMASO
2009

Abstract

In-vivo evaluation of intervertebral kinematics can provide precious information for widespread spinal pathologies such as back pain, whiplash, that still lack of certain diagnoses. Analysis of fluoroscopic sequences screening spine tracts (e.g. lumbar, cervical) during unconstrained patient motion can be used to estimate vertebrae and segmental motion: even if limited, the 2D analysis can be employed to study motion onto sagittal plane. Estimation of vertebral kinematics relies on recognition of vertebrae position and rotation on each radiological frame; this can be achieved identifying specific feature points or landmarks. Manual selection results tedious and imprecise, automatic vertebrae recognition can be based on image template matching. This study proposes a particular template matching that uses smoothed image derivatives, which enhances main vertebral body outline. Vertebra location result more accurate and precise with respect to previous techniques. Results were tested against known data of a reference calibration model: the root mean square error resulted 0.2 degree for vertebral angles and 0.3 mm for vertebra positions. A further comparison was performed using previous findings obtained by processing real sagittal, lumbar fluoroscopic sequences: the root mean square error resulted 1.2 degree for vertebral angles and 0.8 mm for vertebra positions
9783642038815
File in questo prodotto:
Non ci sono file associati a questo prodotto.

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: http://hdl.handle.net/11588/368737
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus 4
  • ???jsp.display-item.citation.isi??? ND
social impact