An Optimization Approach to Scanning Skin Direct Immunofluorescence Specimens

Asser Samak1, Arnold Wiliem1, Michael Walsh2, Ted Ditchmen2, Arne Troskie2,Sarah Barksdale 2, Rhonda Edwards 2, Anthony Jennings 2, Peter Hobson2, Brian C. Lovell1

1School of ITEE, The University of Queensland, Australia

2Sullivan Nicolaides Pathology, Australia



SkinScan Dataset [description and files]

Matlab Toolbox [description and files]


We propose an optimization framework for developing a fully automated scanning system. The framework allows us to have effective design choices in developing the system as every choice should be based on optimizing the objective function. We apply this framework in developing a fully automated scanning system for skin Direct Immunofluorescence (DIF) test. To that end, we introduce both non-algorithmic and algorithmic methods to optimize the objective function. Whilst the non-algorithmic methods comprise various design choices that could indirectly optimize the framework, the algorithmic methods primarily aim to optimize the objective by computing an optimal scan plan. In this work, we explore two algorithmic methods: (1) a heuristic sliding region approach and (2) a quad-tree approach. To our knowledge, this is one of the first works to describe a fully automated scanning system for skin DIF tests. As such, we propose a novel dataset that is hoped to stimulate the research interest in developing digitizing systems for skin DIF tests. All the described methods were evaluated on this novel dataset. Our scanning system is now part of a digital pathology system which has been fully deployed and routinely used within a pathology laboratory.

SkinScan Dataset

In order to stimulate research interest in this area, we propose a novel dataset, denoted SkinScan dataset, that allows comparisons between various algorithmic methods to generate an efficient scan plan. The SkinScan dataset contains 110 images of low-power magnification scans (5x magnification) of DAPI-stained skin tissue specimens, the scan covers the entire slide-well containing the specimen fluorescing in blue. Each specimen image was captured using a monochrome high dynamic range cooled microscopy camera, which was fitted on a microscope with a plan-Apochromat 5x/0.8 objective lenses and an LED illumination source. We also included a mask for each image generated from detecting the well edges (using hough-transform) to eliminate the fluorescing well edges, and a calibration file for the purpose of mapping to the correct tile size. Along with the images, the dataset also contains a MATLAB toolbox implementing our proposed algorithms and the objective function for evaluation.

Each sample includes the original 5x magnification scan along with a calibration file to identify the location of the scan relative to the microscope's tray, also a mask file used to hide the well edges as they seem to fluoresce as well and hence get picked up as tissue, and finally a masked image where the mask is applied to the original scan.


The SkinScan dataset and features + codes ('Licensed Material') are made available to the scientific community for non-commercial research purposes such as academic research, teaching, scientific publications or personal experimentation. Permission is granted by The University of Queensland to you (the 'Licensee') to use, copy and distribute the Licensed Material in accordance with the following terms and conditions:

  1. Licensee must include a reference to The University of Queensland and the following publication in any published work that makes use of the Licensed Material:

    A.Samak, A.Wiliem, P. Hobson and B.C. Lovell

    Digital Image Computing Techniques and Applications (DICTA), 2015

    Bibtex entry:

    @inproceedings{ skinscan2015optim,

    AUTHOR ={ Asser Samak and Arnold Wiliem and Peter Hobson and Michael Walsh and Ted Ditchmen and Arne Troskie and Sarah Barksdale and Rhonda Edwards and Anthony Jennings and Brian C. Lovell},

    TITLE = {An Optimization Approach to Scanning Skin Direct Immunofluorescence Specimens},

    BOOKTITLE = {Digital Image Computing Techniques and Applications (DICTA), 2015 International Conference on},

    YEAR = {2015}


  2. If Licensee alters the content of the Licensed Material or creates any derivative work, Licensee must include in the altered Licensed Material or derivative work prominent notices to ensure that any recipients know that they are not receiving the original Licensed Material.

  3. Licensee may not use or distribute the Licensed Material or any derivative work for commercial purposes including but not limited to, licensing or selling the Licensed Material or using the Licensed Material for commercial gain.

  4. The Licensed Material is provided 'AS IS', without any express or implied warranties. The University of Queensland does not accept any responsibility for errors or omissions in the Licensed Material.

  5. This original license notice must be retained in all copies or derivatives of the Licensed Material.

  6. All rights not expressly granted to the Licensee are reserved by The University of Queensland.

SkinScan Dataset Download

Images & Masks [download] (65 MB)

Toolbox Download

Matlab code [download] (9 KB)


Further enquiries can be made to {a döt wiliem ät uq döt edu döt au} or {lovell ät itee döt uq döt edu döt au}