Type of the Paper (Article

Ryan Cunningham 1, *, María B Sánchez 1, Greg May 1, and Ian Loram 1, * 6 1 School of Healthcare Science, Manchester Metropolitan University, Manchester, England, UK 7 * Correspondence: [email protected], [email protected] 8 9 Abstract: This paper presents an investigation into the feasibility of using deep learning methods for 10 developing arbitrary full spatial resolution regression analysis of B-mode ultrasound images of human skeletal 11 muscle. In this study we focus on full spatial analysis of muscle fibre orientation, since there is an existing body 12 of work with which to compare results. Previous attempts to automatically estimate fibre orientation from 13 ultrasound are not adequate, often requiring manual region selection, feature engineering, providing low14 resolution estimations (one angle per muscle), and deep muscles are often not attempted. We build upon our 15 previous work in which automatic segmentation was used with plain convolutional neural network (CNN), and 16 deep residual convolutional network (ResNet) architectures, to predict a low-resolution map of fibre orientation 17 in extracted muscle regions. Here, we use deconvolutions and max-pooling (DCNN) to regularise and improve 18 predicted fibre orientation maps for the entire image, including deep muscles, removing the need for automatic 19 segmentation, and we compare our results with the CNN and ResNet, as well as a previously established feature 20 engineering method, on the same task. Dynamic ultrasound images sequences of the calf muscles were acquired 21 (25Hz) from 8 healthy volunteers (4 male, ages: 25-36, median 30). A combination of expert annotation and 22 interpolation/extrapolation provided labels of regional fibre orientation for each image. Neural networks (CNN, 23 ResNet, DCNN) were then trained both with and without dropout using leave one out cross-validation. Our 24 results demonstrated robust estimation of full spatial fibre orientation within approximately 6° error, which was 25 an improvement on previous methods. 26