University of Twente Student Theses

Login

Automated skin prick test reading using digital photography: An optimized correction for the camera perspective and arm curvature for accurate wheal size measurement

Berg, A.M. van den (2024) Automated skin prick test reading using digital photography: An optimized correction for the camera perspective and arm curvature for accurate wheal size measurement.

[img] PDF
4MB
Abstract:Background: A common test for diagnosing allergies is the skin prick test (SPT), which is widely used as it is minimally invasive, fast and low in costs. Reading the test results is performed by manually measuring wheal sizes, however, a major drawback is the inter- and intra-observer variability of the test, leading to inaccurate wheal measurements and lack of reproducibility, indicating there is need for a standardized and reliable SPT reading method. This study focuses on automated SPT reading using digital photography. A new method using a basic placemat with ArUco markers is suggested to optimize a correction for the camera perspective and arm curvature in the images, which is needed to obtain accurate wheal sizes that correspond to the real-world dimensions of the wheals. Methods: The perspective correction was performed using the ArUco markers for a projective image transformation. Its accuracy was tested on photographs taken from different camera angles and validated with a checkerboard as reference object. The curvature correction was performed by recovering the three-dimensional shape of the forearm to a two-dimensional flat surface via image transformation. The performance was evaluated with a checkerboard on six subjects and for different regions on the arm, and compared to the conventional barcode sticker method. For insight in clinical relevance, the location of wheals on the forearm regions was analyzed in 88 SPT photographs and related to the placemat performance. Results: The perspective correction showed a high accuracy for all tested camera angles; a maximum deviation of 1% was measured for 10×10 mm checkerboard squares. The curvature correction using the placemat showed for the inner region of the forearm an acceptable error below 10% of 25 mm2 sized squares; the error increased towards the outer sides of the arm. It was found that most wheals in the SPT appear in the inner regions. The placemat method performed better than the barcode method.Conclusions: This work showed that the correction for camera perspective and arm curvature in photographs of the SPT can be optimized using a basic placemat with ArUco markers. This provides more accurate wheal measurements and therefore contributes to obtaining reproducible and reliable SPT results. The performance of the curvature correction could be further improved and evaluation of the placemat on more subjects with varied ages and arms’ shapes is needed.
Item Type:Essay (Master)
Clients:
Deventer Ziekenhuis, Deventer, The Netherlands
Faculty:TNW: Science and Technology
Subject:44 medicine
Programme:Technical Medicine MSc (60033)
Link to this item:https://purl.utwente.nl/essays/98255
Export this item as:BibTeX
EndNote
HTML Citation
Reference Manager

 

Repository Staff Only: item control page