Automatic kidney segmentation in ultrasound images using subsequent boundary distance regression and pixelwise classification networks

•A fully automatic segment method for clinical ultrasound kidney images.•End-to-end learning of boundary detection and pixelwise classification networks.•Achieved significantly better performance than pixelwise classification networks.•Data-augment improved the segmentation performance. It remains c...

Celý popis

Uložené v:

Podrobná bibliografia
Vydané v:	Medical image analysis Ročník 60; s. 101602
Hlavní autori:	Yin, Shi, Peng, Qinmu, Li, Hongming, Zhang, Zhengqiang, You, Xinge, Fischer, Katherine, Furth, Susan L., Tasian, Gregory E., Fan, Yong
Médium:	Journal Article
Jazyk:	English
Vydavateľské údaje:	Netherlands Elsevier B.V 01.02.2020 Elsevier BV
Predmet:	Aged, 80 and over Artificial neural networks Boundary detection Boundary distance regression Classification Datasets as Topic Deep Learning Feature extraction Female Humans Image classification Image processing Image segmentation Kidney Diseases - diagnostic imaging Kidneys Machine learning Male Neural networks Neural Networks, Computer Pattern Recognition, Automated Pixels Pixelwise classification Regression Ultrasonic imaging Ultrasonography - methods Ultrasound Ultrasound images Boundary detection Ultrasound images Pixelwise classification Boundary distance regression
ISSN:	1361-8415, 1361-8423, 1361-8423
On-line prístup:	Získať plný text
Tagy:	Pridať tag Žiadne tagy, Buďte prvý, kto otaguje tento záznam!

Popis
Shrnutí:	•A fully automatic segment method for clinical ultrasound kidney images.•End-to-end learning of boundary detection and pixelwise classification networks.•Achieved significantly better performance than pixelwise classification networks.•Data-augment improved the segmentation performance. It remains challenging to automatically segment kidneys in clinical ultrasound (US) images due to the kidneys’ varied shapes and image intensity distributions, although semi-automatic methods have achieved promising performance. In this study, we propose subsequent boundary distance regression and pixel classification networks to segment the kidneys automatically. Particularly, we first use deep neural networks pre-trained for classification of natural images to extract high-level image features from US images. These features are used as input to learn kidney boundary distance maps using a boundary distance regression network and the predicted boundary distance maps are classified as kidney pixels or non-kidney pixels using a pixelwise classification network in an end-to-end learning fashion. We also adopted a data-augmentation method based on kidney shape registration to generate enriched training data from a small number of US images with manually segmented kidney labels. Experimental results have demonstrated that our method could automatically segment the kidney with promising performance, significantly better than deep learning-based pixel classification networks. [Display omitted]
Bibliografia:	ObjectType-Article-1 SourceType-Scholarly Journals-1 ObjectType-Feature-2 content type line 14 content type line 23
ISSN:	1361-8415 1361-8423 1361-8423
DOI:	10.1016/j.media.2019.101602