School of
Information Technology and Electrical Engineering

The aim of the digital pathology project is to develop technologies that make pathology workflows much faster and more efficient. The project utilises computer vision, machine learning and pattern recognition methods to create digital pathology scanning systems as well as image-based Computer Aided Diagnostic (CAD) systems. The Advanced Surveillance Group within The University of Queensland, along with its' partners Sullivan Nicolaides Pathology are continuously creating a number of computer image datasets from different pathology areas. This effort is hoped to spur more interest from the community to advance this field.

The list of datasets will continue to grow throughout the life of these projects. Check back here on a regular basis for any new additions.


Funding Acknowledgement

These datasets were funded by the following research grants:

  • ARC Linkages Project - Fusion of Digital Microscopy and Plain Text Reports for Automated Analysis
  • Advanced Queensland Research Fellowship (Early) - Transforming Queensland Health Care via Digital Pathology
  • ARC Linkages Project - Application of Manifold-Based Image Analysis to Identify Subtle Changes in Digitally-Captured Pathology Samples

Featured Projects

Investigating the role of sub-auroral polarization stream electric field in coupled magnetosphere-ionosphere-thermosphere processes

2015-2016: USD25k from AOARD

The proposed project intends to investigate the coupled system of magnetosphere (M), ionosphere (I) and thermosphere (T). Its major goal is to contribute to the better understanding of how the solar wind energy becomes dissipated in the M-I-T system during geomagnetic storms. It aims to investigate how the electromagnetic energy, flowing along the magnetic field lines as Poynting flux, dissipates as Joule heating in the I-T system. We will focus on the contributing role of sub-auroral polarization stream (SAPS) electric (E) field, since the SAPS E-field’s source is in the high-latitude region and SAPS effects are strong in the I-T system causing rapid ion-neutral frictional heating and convecting the thermal plasma towards the daytime cusp and into the polar region. To succeed, 1) Poynting flux time series will be analysed with multi-instrument DMSP latitudinal line plots and 2) GPS TEC maps will be analysed with SuperDARN polar convection maps. CTIPe and TIE-GCM will simulate Joule heating and neutral densities. Investigating this important topic, the contributing roles of SAPS E-field and SAPS effects, the proposed research will add directly to recent AFRL studies focusing on M-I-T system energetics. Utilizing our tailor-made software packages, innovative visualization techniques and comprehensive analysis techniques, observational and numerical results will provide new insights into the coupled M-I-T system and its processes benefitting other research groups and communities as well.


Datasets

Immunology

SNPHEp2 Dataset image reference

SNPHEp2 Dataset

This is a small dataset released before the bigger datasets were constructed. For a bigger dataset, please check the competition dataset.

Further information

separator line

SNPHEp2 dataset image reference

UQ-SNP HEp-2 Cell Competition (UQSNPHEp2) datasets

The UQSNPHEp2 competition datasets were used in various international competitions in 2013, 2014 and 2016. These datasets comprise four subtasks. The main page is still under construction. However the dataset can still be downloaded from the above link.

Download the datasets

Histopathology

SkinScan dataset image reference

SkinScan dataset

The SkinScan dataset is useful for comparing between various algorithmic methods for developing scanning algorithm for Skin DIF tests.

Read more and download the datasets

Microbiology

UQ-SNP Gram Stain Region Selection (UQSNP_GRegion) dataset

UQ-SNP Gram Stain Region Selection (UQSNP_GRegion) dataset

UQSNP_GRegion dataset is constructed for developing fast and accurate region selection that will help the system to choose the best area to scan. This dataset is continuously updated.

Read more and download the datasets

separator line

UQ-SNP Gram Stain Cell Dectection (UQSNP_GCellDetect)

UQ-SNP Gram Stain Cell Detection (UQSNP_GCellDetect)

UQSNP_GCellDetect dataset is constructed for developing cell detection in Gram Stain images.

Read more and download the datasets