Automatic segmentation of calibration device boundaries on ophthalmic optical biometry instruments in optical coherence tomography images

Abstract

Background: As the use of ophthalmic optical biometers in clinical practice has increased, so too has the demand for higher standards in evaluating ocular metrics. Among these metrics, the measurement of the ocular axial length (AL) is a critical task that often requires calibrated devices. Thus, efficient algorithms need to be developed to improve calibration mechanisms and ensure precise measurements. This study aimed to establish an algorithm to determine the pixel heights (PHs) of the boundary vertices of a calibration device in ophthalmic optical biometers to improve the accuracy and repeatability of ocular AL measurements in optical coherence tomography (OCT) images.

Methods: The algorithm employs a series of image morphological processing techniques to delineate the rough boundaries of the calibration device in OCT images. After extracting boundary points, clustering techniques are applied to simplify the data. These clustered boundary points are analyzed for characteristic parameters to refine the boundaries. Finally, a fitting process is used to determine the PHs of the vertices, and performance is then evaluated by tests measuring the repeatability and recognition accuracy of the algorithm.

Results: The proposed method had high repeatability in locating the boundaries of the front and rear sections of the calibration device, and had repeatability deviations not exceeding 0.36 pixels and 0.18 pixels, respectively. Additionally, linear fitting of the computed sub-PH and optical path difference yielded determination coefficients of 0.9998330 and 0.9999863, respectively, indicating an almost perfectly linear relationship. This high degree of linearity demonstrates the exceptional accuracy of the method in locating the calibration device boundaries. These results provide a heuristic approach for the future boundary localization of calibration devices in OCT images for biometers, offering a robust and reliable framework for enhancing calibration precision in ophthalmic optical biometry.

Conclusions: The developed algorithm provides an effective and reliable method for determining the PHs of calibration device vertices in ophthalmic optical biometers. With its high accuracy and repeatability, this valuable tool could enhance ocular AL measurements and improve the overall quality of ocular assessments in clinical practice.

Keywords: Optical coherence tomography (OCT); anterior segment; image segmentation; optical biometer calibration devices; posterior segment.

Figure 1

The algorithm for positioning the anterior segment and posterior segment of the calibration device. I_A, anterior OCT image of the calibration device; I_FM, fitting diagram of upper and lower boundary points in anterior of the calibration device; I_KO, anterior information extraction diagram of calibration device; I_KF, posterior information extraction diagram of calibration device; I_LAB, the final fitting result graph of this algorithm; I_P, posterior OCT image of the calibration device; I_PF, fitting diagram of upper and lower boundary points in posterior of the calibration device; OCT, optical coherence tomography.

Figure 2

The identification results of the anterior and posterior segments of nine representative ophthalmic optical biometer calibration devices with varying sizes. The display sequence from (A-I) is based on their sizes, which range from 9.987 to 34.728 mm.

Figure 3

Linear regression plot showing the relationship between the pixel height of the anterior boundary vertex of the calibration device and the corresponding sample arm optical path length.

Figure 4

Linear regression plot showing the relationship between the pixel height of the posterior boundary vertex of the calibration device and the corresponding sample arm optical path length.