Involuntary eye motion correction in retinal optical coherence tomography: Hardware or software solution?

Abstract

In this paper, we review state-of-the-art techniques to correct eye motion artifacts in Optical Coherence Tomography (OCT) imaging. The methods for eye motion artifact reduction can be categorized into two major classes: (1) hardware-based techniques and (2) software-based techniques. In the first class, additional hardware is mounted onto the OCT scanner to gather information about the eye motion patterns during OCT data acquisition. This information is later processed and applied to the OCT data for creating an anatomically correct representation of the retina, either in an offline or online manner. In software based techniques, the motion patterns are approximated either by comparing the acquired data to a reference image, or by considering some prior assumptions about the nature of the eye motion. Careful investigations done on the most common methods in the field provides invaluable insight regarding future directions of the research in this area. The challenge in hardware-based techniques lies in the implementation aspects of particular devices. However, the results of these techniques are superior to those obtained from software-based techniques because they are capable of capturing secondary data related to eye motion during OCT acquisition. Software-based techniques on the other hand, achieve moderate success and their performance is highly dependent on the quality of the OCT data in terms of the amount of motion artifacts contained in them. However, they are still relevant to the field since they are the sole class of techniques with the ability to be applied to legacy data acquired using systems that do not have extra hardware to track eye motion.

Keywords: Motion artifact correction; Optical Coherence Tomography (OCT); Retina.

Figure 1

Sample OCT volume with the corresponding A-scan axis (AA), fast-scanning axis (FA) and slow-scanning axis (SA)

Figure 2

An example of the work by Tao et al. (2010) for hardware-based eye motion correction using interlaced SECSLO-SDOCT. (a) A region of interest in the SECSLO frame, (b) en face view of the SD-OCT volume without correction, (c) en face view of the SD-OCT volume after correction, (d) corrected en face view re-sampled to a regular image grid. [Reprinted with permission]

Figure 3

An example of the work by Vienola et al. (2012) for online motion compensation: (A) and (C) without tracking and (B) and (D) with tracking. The online implementation enables correction of large saccades and re-scanning of the corrupted regions (D). [Reprinted with permission]

Figure 4

An examples of eye motion artifacts in phase-resolved OCT angiography of the retina in the work of Braaf et al. (2013). The online implementation enables correction of large saccades and re-scanning of the corrupted regions (D). [Reprinted with permission]

Figure 5

An example of the work by LaRocca et al. (2013b). The left column shows the two en face view of orthogonal scans without motion correction. The middle row displays the SLO-based motion corrected en face views without resampling. The right column contains the results of the motion correction after resampling. The images in the bottom row are the results of combining the two en face views. [Reprinted with permission]

Figure 6

Axial motion correction by use of three orthogonal B-scans located at the beginning, middle and at the end of the imaged volume (Potsaid et al. (2008)). Using ultra-high speed OCT devices, the true curvature of the retina can be reconstructed with higher degrees of reliability. [Reprinted with permission]

Figure 7

An example of the result of the work by Antony et al. (2011) for axial motion correction of OCT volume scans. f and s subscripts represent the fast-scanning and slow-scanning axes, respectively. The technique will result in a flattened volume. But with use of orthogonal volume scans, the true curvature of the retina can be reconstructed. [Reprinted with permission]

Figure 8

An example of the work by Hendargo et al. (2013). Two X-fast and two Y-fast datasets were acquired. The original SVPs for each of the three main vessel layers are shown in the left four columns. The fifth column shows the results of image registration for each of the three layers. Each image covers a 2.5 2.5 mm scan area. Motion artifacts are removed and visualization of the vasculature is enhanced in the registered images. A color encoded depth image is shown on the right, combining information from the registered images of the three vessel layers. Red indicates more superficial vessels while blue indicates deeper vessels. [Reprinted with permission]

Figure 9

An example of the work by Kraus et al. (2014). The first row represents the uncorrected blood vessel maps from the two input orthogonal volume scans as well as their difference map. The second row is the result of applying the first stage of the algorithm. The third row is the result of the final stage of the proposed algorithm. [Reprinted with permission]