Computer-aided detection system for nerve identification using ultrasound images: A comparative study

Oussama Hadjerci (1) , Adel Hafiane (1) , Pierre Vieyres (1) , Donatello Conte (2) , Pascal Makris (2) , Alain Delbos (3)

1 PRISME - Laboratoire Pluridisciplinaire de Recherche en Ingénierie des Systèmes, Mécanique et Energétique
2 LIFAT - Laboratoire d'Informatique Fondamentale et Appliquée de Tours
3 Clinique Médipôle Garonne

Ultrasound-Guided Regional Anesthesia (UGRA) has been gaining importance in the last few years, offering numerous advantages over alternative methods of nerve localization (neurostimulation or paraesthesia). However, nerve detection is one of the most difficult tasks that anesthetists can encounter in the UGRA procedure. The context of the present work is to provide practitioners with a computer-aided system to facilitate the practice of UGRA. However, automatic detection and segmentation in US images is still a challenging problem in many medical applications. This paper addresses two main issues, first proposing an efficient framework for nerve detection and segmentation, and second, reviewing literature methods and evaluating their performance for this new application. The proposed system consists of four main stages: (1) despeckling filter, (2) feature extraction, (3) feature selection, (4) classification and segmentation. A comparative study was performed in each of these stages to measure their influence over the whole system. Sonographic videos were acquired with the same ultrasound machine in real conditions from 19 volunteer patients. Evaluation was designed to cover two important aspects: measure the effect of training set size, and evaluate consistency using a cross-validation technique. The results obtained were significant and indicated which method was better for a nerve detection system. The proposed scheme achieved high scores (i.e.80% on average of 1900 tested images), demonstrating its validity.

Mots clés: Feature selection; Despeckling; Feature extraction; Computer-aided detection; Ultrasound image processing; Machine learning; Classification

Domaines: Imagerie médicale; Traitement des images [eess.IV]; Traitement du signal et de l'image [eess.SP]

Oussama Hadjerci, Adel Hafiane, Pierre Vieyres, Donatello Conte, Pascal Makris, et al.. Computer-aided detection system for nerve identification using ultrasound images: A comparative study. Informatics in Medicine Unlocked, 2016, 3, pp.29-43. ⟨10.1016/j.imu.2016.06.003⟩. ⟨hal-01405133⟩