IFSS-Net: Interactive Few-Shot Siamese Network for Faster Muscle Segmentation and Propagation in Volumetric Ultrasound

• Published in: IEEE transactions on medical imaging Vol. 40; no. 10; pp. 2615 - 2628
• Main Authors: Chanti, Dawood Al, Duque, Vanessa Gonzalez, Crouzier, Marion, Nordez, Antoine, Lacourpaille, Lilian, Mateus, Diana
• Format: Journal Article
• Language: English
• Published: United States IEEE 01.10.2021; The Institute of Electrical and Electronics Engineers, Inc. (IEEE); Institute of Electrical and Electronics Engineers
• Subjects: 3D ultrasound; Annotations; Artificial neural networks; Coders; Computer Science; Computer Vision and Pattern Recognition; Encoders-Decoders; few-shot annotation; Image Processing; Image segmentation; Learning; Long short-term memory; Machine Learning; mask propagation; Muscles; Objective function; Propagation; pseudo-labeling; segmentation; Statistics; Switches; Task analysis; Three-dimensional displays; Training; Two dimensional displays; Ultrasonic imaging; Ultrasound; Few-Shot annotation; Mask propagation; Segmentation; 3D Ultrasound; Pseudo labelling
• ISSN: 0278-0062; 1558-254X; 1558-254X
• DOI: 10.1109/TMI.2021.3058303

We present an accurate, fast and efficient method for segmentation and muscle mask propagation in 3D freehand ultrasound data, towards accurate volume quantification. A deep Siamese 3D Encoder-Decoder network that captures the evolution of the muscle appearance and shape for contiguous slices is deployed. We use it to propagate a reference mask annotated by a clinical expert. To handle longer changes of the muscle shape over the entire volume and to provide an accurate propagation, we devise a Bidirectional Long Short Term Memory module. Also, to train our model with a minimal amount of training samples, we propose a strategy combining learning from few annotated 2D ultrasound slices with sequential pseudo-labeling of the unannotated slices. We introduce a decremental update of the objective function to guide the model convergence in the absence of large amounts of annotated data. After training with a few volumes, the decremental update strategy switches from a weak supervised training to a few-shot setting. Finally, to handle the class-imbalance between foreground and background muscle pixels, we propose a parametric Tversky loss function that learns to penalize adaptively the false positives and the false negatives. We validate our approach for the segmentation, label propagation, and volume computation of the three low-limb muscles on a dataset of 61600 images from 44 subjects. We achieve a Dice score coefficient of over 95% and a volumetric error of 1.6035 ± 0.587%.