Annotation Cost Minimization for Ultrasound Image Segmentation using Cross-domain Transfer Learning.

Deep learning techniques can help minimize inter-physician analysis variability and the medical expert workloads, thereby enabling more accurate diagnoses. However, their implementation requires large-scale annotated dataset whose acquisition incurs heavy time and human-expertise costs. Hence, to significantly minimize the annotation cost, this study presents a novel framework that enables the deployment of deep learning methods in ultrasound (US) image segmentation requiring only very limited manually annotated samples. We propose SegMix, a fast and efficient approach that exploits a segment-paste-blend concept to generate large number of annotated samples based on a few manually acquired labels. Besides, a series of US-specific augmentation strategies built upon image enhancement algorithms are introduced to make maximum use of the available limited number of manually delineated images. The feasibility of the proposed framework is validated on the left ventricle (LV) segmentation and fetal head (FH) segmentation tasks, respectively. Experimental results demonstrate that using only 10 manually annotated images, the proposed framework can achieve a Dice and JI of 82.61% and 83.92%, and 88.42% and 89.27% for LV segmentation and FH segmentation, respectively. Compared with training using the entire training set, there is over 98% of annotation cost reduction while achieving comparable segmentation performance. This indicates that the proposed framework enables satisfactory deep leaning performance when very limited number of annotated samples is available. Therefore, we believe that it can be a reliable solution for annotation cost reduction in medical image analysis.

US-Net: A lightweight network for simultaneous speckle suppression and texture enhancement in ultrasound images.

BACKGROUND: Numerous traditional filtering approaches and deep learning-based methods have been proposed to improve the quality of ultrasound (US) image data. However, their results tend to suffer from over-smoothing and loss of texture and fine details. Moreover, they perform poorly on images with different degradation levels and mainly focus on speckle reduction, even though texture and fine detail enhancement are of crucial importance in clinical diagnosis. METHODS: We propose an end-to-end framework termed US-Net for simultaneous speckle suppression and texture enhancement in US images. The architecture of US-Net is inspired by U-Net, whereby a feature refinement attention block (FRAB) is introduced to enable an effective learning of multi-level and multi-contextual representative features. Specifically, FRAB aims to emphasize high-frequency image information, which helps boost the restoration and preservation of fine-grained and textural details. Furthermore, our proposed US-Net is trained essentially with real US image data, whereby real US images embedded with simulated multi-level speckle noise are used as an auxiliary training set. RESULTS: Extensive quantitative and qualitative experiments indicate that although trained with only one US image data type, our proposed US-Net is capable of restoring images acquired from different body parts and scanning settings with different degradation levels, while exhibiting favorable performance against state-of-the-art image enhancement approaches. Furthermore, utilizing our proposed US-Net as a pre-processing stage for COVID-19 diagnosis results in a gain of 3.6% in diagnostic accuracy. CONCLUSIONS: The proposed framework can help improve the accuracy of ultrasound diagnosis.

An efficient annotated data generation method for echocardiographic image segmentation.

BACKGROUND: In recent years, deep learning techniques have demonstrated promising performances in echocardiography (echo) data segmentation, which constitutes a critical step in the diagnosis and prognosis of cardiovascular diseases (CVDs). However, their successful implementation requires large number and high-quality annotated samples, whose acquisition is arduous and expertise-demanding. To this end, this study aims at circumventing the tedious, time-consuming and expertise-demanding data annotation involved in deep learning-based echo data segmentation. METHODS: We propose a two-phase framework for fast generation of annotated echo data needed for implementing intelligent cardiac structure segmentation systems. First, multi-size and multi-orientation cardiac structures are simulated leveraging polynomial fitting method. Second, the obtained cardiac structures are embedded onto curated endoscopic ultrasound images using Fourier Transform algorithm, resulting in pairs of annotated samples. The practical significance of the proposed framework is validated through using the generated realistic annotated images as auxiliary dataset to pretrain deep learning models for automatic segmentation of left ventricle and left ventricle wall in real echo data, respectively. RESULTS: Extensive experimental analyses indicate that compared with training from scratch, fine-tuning after pretraining with the generated dataset always results in significant performance improvement whereby the improvement margins in terms of Dice and IoU can reach 12.9% and 7.74%, respectively. CONCLUSION: The proposed framework has great potential to overcome the shortage of labeled data hampering the deployment of deep learning approaches in echo data analysis.