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


Current browse context:


Change to browse by:

References & Citations


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

Quantitative Biology > Quantitative Methods

Title: Quantitative Imaging Principles Improves Medical Image Learning

Abstract: Fundamental differences between natural and medical images have recently favored the use of self-supervised learning (SSL) over ImageNet transfer learning for medical image applications. Differences between image types are primarily due to the imaging modality and medical images utilize a wide range of physics based techniques while natural images are captured using only visible light. While many have demonstrated that SSL on medical images has resulted in better downstream task performance, our work suggests that more performance can be gained. The scientific principles which are used to acquire medical images are not often considered when constructing learning problems. For this reason, we propose incorporating quantitative imaging principles during generative SSL to improve image quality and quantitative biological accuracy. We show that this training schema results in better starting states for downstream supervised training on limited data. Our model also generates images that validate on clinical quantitative analysis software.
Subjects: Quantitative Methods (q-bio.QM); Computer Vision and Pattern Recognition (cs.CV); Image and Video Processing (eess.IV)
Cite as: arXiv:2206.06663 [q-bio.QM]
  (or arXiv:2206.06663v3 [q-bio.QM] for this version)

Submission history

From: Lambert Leong [view email]
[v1] Tue, 14 Jun 2022 07:51:49 GMT (4869kb)
[v2] Fri, 1 Jul 2022 22:10:11 GMT (4868kb)
[v3] Mon, 11 Jul 2022 22:12:34 GMT (4869kb)

Link back to: arXiv, form interface, contact.