We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

eess.IV

Change to browse by:

References & Citations

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Electrical Engineering and Systems Science > Image and Video Processing

Title: Chest X-ray lung and heart segmentation based on minimal training sets

Authors: Balázs Maga
Abstract: As the COVID-19 pandemic aggravated the excessive workload of doctors globally, the demand for computer aided methods in medical imaging analysis increased even further. Such tools can result in more robust diagnostic pipelines which are less prone to human errors. In our paper, we present a deep neural network to which we refer to as Attention BCDU-Net, and apply it to the task of lung and heart segmentation from chest X-ray (CXR) images, a basic but ardous step in the diagnostic pipeline, for instance for the detection of cardiomegaly. We show that the fine-tuned model exceeds previous state-of-the-art results, reaching $98.1\pm 0.1\%$ Dice score and $95.2\pm 0.1\%$ IoU score on the dataset of Japanese Society of Radiological Technology (JSRT). Besides that, we demonstrate the relative simplicity of the task by attaining surprisingly strong results with training sets of size 10 and 20: in terms of Dice score, $97.0\pm 0.8\%$ and $97.3\pm 0.5$, respectively, while in terms of IoU score, $92.2\pm 1.2\%$ and $93.3\pm 0.4\%$, respectively. To achieve these scores, we capitalize on the mixup augmentation technique, which yields a remarkable gain above $4\%$ IoU score in the size 10 setup.
Comments: Preprint. arXiv admin note: text overlap with arXiv:2003.10304
Subjects: Image and Video Processing (eess.IV); Computer Vision and Pattern Recognition (cs.CV)
Cite as: arXiv:2101.08309 [eess.IV]
  (or arXiv:2101.08309v1 [eess.IV] for this version)

Submission history

From: Balázs Maga [view email]
[v1] Wed, 20 Jan 2021 20:24:35 GMT (2540kb,D)

Link back to: arXiv, form interface, contact.