Local Phase Tensor Features for 3-D Ultrasound to Statistical Shape+Pose Spine Model Registration

Abstract

Most conventional spine interventions are performed under X-ray fluoroscopy guidance. In recent years, there has been a growing interest to develop nonionizing imaging alternatives to guide these procedures. Ultrasound guidance has emerged as a leading alternative. However, a challenging problem is automatic identification of the spinal anatomy in ultrasound data. In this paper, we propose a local phase-based bone feature enhancement technique that can robustly identify the spine surface in ultrasound images. The local phase information is obtained using a gradient energy tensor filter. This information is used to construct local phase tensors in ultrasound images, which highlight the spine surface. We show that our proposed approach results in a more distinct enhancement of the bone surfaces compared to recently proposed techniques based on monogenic scale-space filters and logarithmic Gabor filters. We also demonstrate that registration accuracy of a statistical shape+pose model of the spine to 3-D ultrasound images can be significantly improved, using the proposed method, compared to those obtained using monogenic scale-space filters and logarithmic Gabor filters.

Segmentation of the spinous process and its acoustic shadow in vertebral ultrasound images

“…When the convex hull is computed, the size of this overextension is reduced but it is not entirely removed. To address this problem, the use of a shape template for the acoustic shadow, designed from the input of several experts (Tran and Rohling, 2010) or a statistical atlas (Khallaghi et al, 2010;Hacihaliloglu et al, 2014), could be a solution.…”

“…To improve accuracy further, our automatic segmentation method can easily be extended to utilize new image features. For instance, Hacihaliloglu et al's phase symmetry features (Hacihaliloglu et al, 2009) might advantageously be replaced by their more recently developed local phase tensor features (Hacihaliloglu et al, 2014). Future work will also involve optimizing the image acquisition parameters and protocol in order to capture other anatomical structures like the transverse processes, the laminae and the vertebral bodies, which will be integrated as new classes in the segmentation algorithm.…”

“…Khallaghi et al (2010) optimize the similarity between the ultrasound data and synthetic ultrasound images generated from the 3D spine model using a simulation method. Hacihaliloglu et al (2014) propose instead to match bone surfaces from the model to bone surfaces enhanced using local phase tensor features in the ultrasound image. Dense segmentation of the ultrasound images, as proposed in this paper, and segmentation of acoustic shadow in particular, could certainly be exploited as part of such modelbased registration strategies to utilize only meaningful correspondences, as proposed by Coupé et al (2012) in the context of image-guided neurosurgery.…”

Analytic description of the image to patient torso registration problem in image guided interventions

“…The same approach can be performed with other kind of intra-operative imaging device e.g. an ultrasound to acquire clearly distinguishable internal anatomical references [47] or directly the entire shape of the structure of interest [48] .…”

Ultrasound imaging and segmentation of bone surfaces: A review

“…The use of local phase image information in the context of segmentation and/or enhancement of bone structures from US data was first proposed by Hacihaliloglu et al . 52 and extensively investigated in their consecutive publications 53–56 . In their initial work, a 2D Log-Gabor filter was used to extract the image phase information which was combined into a feature descriptor named phase symmetry (PS) in order to extract bone features 52 .…”