Phase-based probabilistic active contour for nerve detection in ultrasound images for regional anesthesia Adel Hafiane a,n , Pierre Vieyres b , Alain Delbos c a INSA Centre Val de Loire, Univ. Orléans, PRISME EA 4229, 88 boulevard Lahitolle, F-18022 Bourges, France b IUT de Bourges, Univ. Orléans, PRISME EA 4229, 63 Avenue de Lattre de Tassigny, F-18020 Bourges, France c Clinique Medipole Garonne, 45 rue de Gironis CS 13624, F-31036 Toulouse, France article info Article history: Received 21 June 2012 Accepted 2 June 2014 Keywords: Ultrasound images Medical image processing Active contours Probabilistic learning Monogenic signal Regional anesthesia abstract Ultrasound guided regional anesthesia (UGRA) is steadily growing in popularity, owing to advances in ultrasound imaging technology and the advantages that this technique presents for safety and efficiency. The aim of this work is to assist anaesthetists during the UGRA procedure by automatically detecting the nerve blocks in the ultrasound images. The main disadvantage of ultrasound images is the poor quality of the images, which are also affected by the speckle noise. Moreover, the nerve structure is not salient amid the other tissues, which makes its detection a challenging problem. In this paper we propose a new method to tackle the problem of nerve zone detection in ultrasound images. The method consists in a combination of three approaches: probabilistic, edge phase information and active contours. The gradient vector flow (GVF) is adopted as an edge-based active contour. The phase analysis of the monogenic signal is used to provide reliable edges for the GVF. Then, a learned probabilistic model reduces the false positives and increases the likelihood energy term of the target region. It yields a new external force field that attracts the active contour toward the desired region of interest. The proposed scheme has been applied to sciatic nerve regions. The qualitative and quantitative evaluations show a high accuracy and a significant improvement in performance. & 2014 Elsevier Ltd. All rights reserved. 1. Introduction Regional anesthesia (RA) is a technique used to inhibit sensa- tion in a particular region of human body, such as an arm or a leg. It consists in distributing the anesthetic product around the nerve structures to eliminate sensitivity and mobility. The main advan- tage of regional anesthesia is that it immobilizes a selected region of the body while having little effect on other parts, such as respiratory functions. This enables fast recovery and reduces the side effects when compared to general anesthesia. The benefit of RA is particularly pronounced for pain management during sur- gery and other medical procedures [1]. The key requirement for successful regional anesthesia is localization of the nerve block in order to ensure optimal distribution of the local anesthetic around nerve structures [2]. The conventional method of nerve localiza- tion uses nerve stimulation, which presents a high risk of nerve trauma. The use of ultrasound (US) imaging for RA is steadily increasing: its advantages over conventional techniques are significant, since it allows direct ultrasonographic guidance of the needle to the nerve zone [3–5]. The nerve structure includes several types of tissues. A single nerve fiber is surrounded by endoneurium. A group of nerve fibers forms a nerve fascicle. Each nerve fascicle is surrounded by perineurium. Nerve fascicles together form a nerve which is surrounded by epineurium. Nerve tissues have different behaviors under ultrasound waves. Nerve fibers themselves do not reflect any ultrasound (hypoechoic) so they appear dark. Only the con- nective tissue surrounding the nerve fibers, the fascicles and the nerve (Epineurium) reflects ultrasound (hyperechoic) and thus appears bright [4]. Fig. 1 shows an example of the nerve anatomy and its cross-section visualized by ultrasound imaging. The use of ultrasound guided regional anesthesia (UGRA) in daily clinical practice requires a high degree of training and practical skills to identify the nerve block and steer the needle to it [2]. This can limit the development and the generalization of the practice of UGRA. There are two critical steps in UGRA: the recognition of anatomical structures (i.e. target nerves, vascular structures, etc.) and needle tracking during insertion. The aim is to develop an assistance system that can handle these critical issues, hoping to spark a shift towards easier practice of regional anesthesia under sonographic guidance. This work focuses on Contents lists available at ScienceDirect journal homepage: www.elsevier.com/locate/cbm Computers in Biology and Medicine http://dx.doi.org/10.1016/j.compbiomed.2014.06.001 0010-4825/& 2014 Elsevier Ltd. All rights reserved. n Corresponding author. E-mail address: adel.hafiane@insa-cvl.fr (A. Hafiane). Computers in Biology and Medicine 52 (2014) 88–95