XJTU

Artificial Intelligence in Pathology

Transition from conventional to digital pathology -- With virtual slides, computational algorithms can be applied to analyze tissue structures at different levels: cellular, inter-cellular, cell architecture, texture of the tissue, etc. Connecting pathology to other areas in medicine is also possible.

Overview of artificial intelligence in pathology

Artificial intelligence has had successful applications in fields like computer vision and natural language processing. Behind these successes was the availability of large-scale annotated datasets. In the past decade, digitized histopathological images have gained quantity and accessibility so that application of machine learning techniques are feasible. The potential of the technology is far from being fully exploited in pathology. Our group, Biomedical Semantic Understanding Group, in Xi'an Jiaotong University, is working on facilitating transition of conventional to digital pathology. We believe that this can be done by

  • Standardized data exchanging standard
  • Open source software and tools for digital pathology
  • Large-scale datasets with annotations
  • We describe our work in this site to promote development in digital pathology.

    Since 2018, we have been developing and maintaining OpenHI, a digital pathology platform. This will facilitate transition from conventional to digital pathology in several ways: establish easy-to-use tools for pathologists and data scientists to work with multi-resolution images, reduce efforts in large-scale histopathological image annotation, and integrate semantic standards into annotated pathological database. Other than the platform, our group organize workshops as a forum for researchers in digital pathology. So far, AIPath2019 was organized in November 2019 in conjunction with BIBM2019

    To get latest updates from us, please sign up for our newsletter (Frequency: less than once a month).