The primary objective of the study is to automate and decentralise the Skin Prick Test, with a focus on the automatic detection and segmentation of wheals. Besides, the possibilities of implementing a DPM that outputs the probability on a certain allergy will be investigated. The following research question is defined for the completion of the study objective:
To what extend can the Skin Prick Test be automated and decentralised, containing a probability-based output for allergy diagnosis by children with a presumably present allergy?
The secondary objectives of the study include (I) an extensive study on image pre-processing possibilities prior to wheal detection and segmentation with respect to different colour spaces and texture features,(II) comparison of an AI-driven segmentation method and a Computer Vision-based segmentation method and (III) parameter identification and analysis for the DPM.
The subsidiary questions are:
1. Which pre-processing steps are valuable prior to wheal segmentation?
2. What is the best automatic wheal detection and segmentation algorithm?
3. What patient information and which SPT parameters influence the patient's outcome and in what manner?
4. What is the best DPM to predict the probability of a certain allergy based on the automated Skin Prick Test result?