Joint Segmentation and Landmark Localization of Fetal Femur in Ultrasound Volumes

Abstract

Volumetric ultrasound has great potentials in promoting prenatal examinations. Automated solutions are highly desired to efficiently and effectively analyze the massive volumes. Segmentation and landmark localization are two key techniques in making the quantitative evaluation of prenatal ultrasound volumes available in clinic. However, both tasks are non-trivial when considering the poor image quality, boundary ambiguity and anatomical variations in volumetric ultrasound. In this paper, we propose an effective framework for simultaneous segmentation and landmark localization in prenatal ultrasound volumes. The proposed framework has two branches where informative cues of segmentation and landmark localization can be propagated bidirectionally to benefit both tasks. As landmark localization tends to suffer from false positives, we propose a distance based loss to suppress the noise and thus enhance the localization map and in turn the segmentation. Finally, we further leverage an adversarial module to emphasize the correspondence between segmentation and landmark localization. Extensively validated on a volumetric ultrasound dataset of fetal femur, our proposed framework proves to be a promising solution to facilitate the interpretation of prenatal ultrasound volumes.

A Review on Deep-Learning Algorithms for Fetal Ultrasound-Image Analysis

“…However, it is hard to be evaluated through US, due to the difficulty in locating the tips of the femur, boundary deficiency and ambiguity for tissues' low contrasts, and variation of pose, shape and size of this structure. In [107], the authors develop a unified framework for simultaneous segmentation and landmark localization of fetal femur in prenatal US volumes: fetal femur ROI is first identified through a U-Net model, then segmentation and landmark localization branches receive the common features of ROI extracted by the shared layers and generate task-specific descriptors. The method tested on 20 US volumes reaches DSC, IoU , HD and ED equal to 0.91, 0.83, 4.08 mm and 0.87 mm, respectively.…”

Shape Constraints in Deep Learning for Robust 2D Echocardiography Analysis

“…As for landmark detection, we first penalise on the squared error of landmark heat-map (L2 loss: L l2 ). In order to avoid landmark overlapping on different output layers, we regularise the centre distance loss L CD of different landmark heat-maps as proposed in [15]:…”

3D Bounding Box Detection in Volumetric Medical Image Data: A Systematic Literature Review

“…A common approach is to use a single 2D CNN or a combination (2.5D) of several (usually three) 2D CNNs for slice wise detection in either one or all three orthogonal viewing plane directions. A single 2D CNN can be implemented to analyze exactly one of the three image plane stacks [27], [28], [29], [30]. Adjacent slices as additional channels [31] or dimensions [32] help to capture contextual information.…”

“…First, the entire image is viewed to roughly locate one or more targets. The resulting sub-optimal segmentation is then utilized to place a BB around the area of interest [32], [33], [34], [29], [5], [31], [19], [20], [39]. Similar to a coarse-segmentation approach, H. Roth et al (2018) [38] implement a 2D pixel-wise probability detection in every image plane direction to obtain confidence heatmaps, which are then used to generate a 3D BB.…”