期刊
IEEE TRANSACTIONS ON IMAGE PROCESSING
卷 29, 期 -, 页码 6667-6679出版社
IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
DOI: 10.1109/TIP.2020.2992893
关键词
Image segmentation; Computed tomography; Soil; Image resolution; Decoding; Three-dimensional displays; Biomedical imaging; X-ray computed tomography; image segmentation; deep learning; root system analysis; plant phenotyping
资金
- U.S. Department of Energy via the ARPA-e ROOTS Project Low Cost X-Ray CT System for in-situ Imaging of Roots
- European Research Council (ERC)
- University of Nottingham
- Future Food Beacon of Excellence
We address the complex problem of reliably segmenting root structure from soil in X-ray Computed Tomography (CT) images. We utilise a deep learning approach, and propose a state-of-the-art multi-resolution architecture based on encoder-decoders. While previous work in encoder-decoders implies the use of multiple resolutions simply by downsampling and upsampling images, we make this process explicit, with branches of the network tasked separately with obtaining local high-resolution segmentation, and wider low-resolution contextual information. The complete network is a memory efficient implementation that is still able to resolve small root detail in large volumetric images. We compare against a number of different encoder-decoder based architectures from the literature, as well as a popular existing image analysis tool designed for root CT segmentation. We show qualitatively and quantitatively that a multi-resolution approach offers substantial accuracy improvements over a both a small receptive field size in a deep network, or a larger receptive field in a shallower network. We then further improve performance using an incremental learning approach, in which failures in the original network are used to generate harder negative training examples. Our proposed method requires no user interaction, is fully automatic, and identifies large and fine root material throughout the whole volume.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据