Automatic diagnosis of strabismus in digital videos through cover test

doi:10.1016/j.cmpb.2017.01.002

Computer Methods and Programs in Biomedicine

Volume 140, March 2017, Pages 295-305

https://doi.org/10.1016/j.cmpb.2017.01.002 Get rights and content

Highlights

•
This work investigates computational method for automatic diagnose of strabismus.
•
The strabismus was detected in digital video through the Cover Test.
•
The method uses image and video processing and requires only a digital camera and a regular computer.
•
The method achieved 87% of accuracy in diagnosing strabismus.
•
The overall average error was lower than 1Δ, and an average error of 2.6Δ in deviation measure.

Abstract

Background and Objective: Medical image processing can contribute to the detection and diagnosis of human body anomalies, and it represents an important tool to assist in minimizing the degree of uncertainty of any diagnosis, while providing specialists with an additional source of diagnostic information. Strabismus is an anomaly that affects approximately 4% of the population. Strabismus modifies vision such that the eyes do not properly align, influencing binocular vision and depth perception. Additionally, it results in aesthetic problems, which can be reversed at any age. However, the use of low cost computational resources to assist in the diagnosis and treatment of strabismus is not yet widely available. This work presents a computational methodology to automatically diagnose strabismus through digital videos featuring a cover test using only a workstation computer to process these videos.

Methods: The method proposed was validated in patients with exotropia and consists of eight steps: (1) acquisition, (2) detection of the region surrounding the eyes, (3) identification of the location of the pupil, (4) identification of the location of the limbus, (5) eye movement tracking, (6) detection of the occluder, (7) identification of evidence of the presence of strabismus, and (8) diagnosis.

Results: To detect the presence of strabismus, the proposed method achieved a specificity value of 100%, and (2) a sensitivity value of 80%, with 93.33% accuracy in diagnosis of patients with extropia. This procedure was recognized to diagnose strabismus with an accuracy value of 87%, while acknowledging measures lower than 1Δ, and an average error in the deviation measure of 2.57Δ.

Conclusions: We demonstrated the feasibility of using computational resources based on image processing techniques to achieve success in diagnosing strabismus by using the cover test. Despite the promising results the proposed method must be validated in a greater volume of video including other types of strabismus.

Introduction

Strabismus is a condition in which the line of sight of one eye does not fixate on the object upon which both eyes are focused [1]. That is, while one of the eyes focuses on the object, the other eye focuses toward another direction, generating a misalignment of the ocular muscles in relation to the eye focus at the frontal point. This anomaly can be classified according to the focus direction in which the deviated eye points, which can be either manifest or latent. The former is identified by simply looking at the patient; the latter requires tests to obtain the diagnosis.

Studies show that 65% of people with strabismus developed this anomaly by the age of 3 years [2], [3]. Several studies elaborate upon the importance of early treatment of strabismus in order to reduce undesirable social and emotional consequences to a person’s life, such as depression, low self-esteem, bullying, and relationship issues [4], [5], [6], [7], [8], [9], [10]. Children usually express negative feelings towards the condition around the age of 6 years by expressing signs of hostility and dislike [8]. In [5], subsequent to strabismus surgery, 61% of children younger than 4 years exhibit increased visual contact, and 55% of children between 4 and 6 years exhibit improvement of their self-esteem. Research has also identified that strabismus has a negative impact on the quality of life of adult patients. These impacts can cause psychological issues and the inability to perform daily activities caused by binocular diplopia¹ related to strabismus [11].

Several tests are usually required to diagnose strabismus. Among these tests, the most common test is the cover test, qualified by a subtype: (a) the prism cover test (PCT) known as the ‘golden standard’ by specialists [12], (b) the unilateral cover test used to detect strabismus [13] and (c) the alternate cover test used to diagnose strabismus [14]. These tests are included within the category of ocular motility tests that are used for the detection and diagnosis of strabismus, either manifest or not. When performing the cover test, one of the eyes is occluded by an opaque object. Next, the occluder is shifted from one eye to the other, alternating back and forth, and the direction and estimated magnitude of eye movement are observed [15].

Currently, the number of strabismus specialists located in urban areas is relatively small, and it is difficult to find a strabismus specialist in a suburban area. The availability and use of high technology tools and resources to assist in ophthalmology diagnostics is limited within the subspecialty of strabismus, despite publication results of some studies in recent years.

Some of these studies are based entirely on manifest strabismus and are conducted by the use of tools that are difficult to obtain and utilize by non-specialized professionals in strabismus or professionals whose understanding of the requisite fully automated technology is limited. The purpose of this work is to propose an automated methodology that is based on the alternate cover test to solve the shortcomings of the aforementioned studies. The tests are conducted using computational resources and provide a second opinion to be analyzed by strabismus specialists.

In addition, this work contains additional sections that address the objective and results of the proposed methodology. In Section 2, we briefly present some studies related to strabismus diagnosis using computational resources. In Section 3, we describe the eight stages of the methodology, from the acquisition of videos to the diagnosis of strabismus. In Section 4 and Section 5, we present and discuss the results achieved by this methodology. Finally, in Section 6, we present the conclusions, contributions, and suggestions for future work.

Section snippets

Related work

Although the use of computational resources to aid ophthalmology specialists is relatively recent, studies that present both tools and methodologies already exist. This section introduces research, equipment, and related work that support the detection and diagnosis of strabismus.

Regarding the equipment used for the detection and diagnosis of eye pathologies, we cite Eye Tracking [16], which is used in ocular motility research labs to measure deviations and eye movement, and Electronic

Proposed methodology

To achieve a reliable and sensitive diagnosis, it is necessary to identify, in each frame, the extent to which the eye involuntarily moves when applying the cover test. To satisfy this requirement, the proposed methodology is organized into eight stages, as illustrated in Fig. 1. The image of the face was disguised by blurring it at the top and bottom regions to preserve the patient’s identity. The numbers on the image help to guide the sequence of stages.

Our procedure consists of the following

Results

This section presents the results achieved by the proposed method. We consider the collection of the patient’s videos in all execution stages of the methodology, from the reduction of eye region stage to the strabismus diagnosis stage. At the conclusion of the process, the results achieved are validated by comparing them with the diagnosis provided by specialists.

Discussion

This section discusses the results achieved by the proposed method. In detection of eye region and pupil location stage (Section 4.1) the methodology failed in six cases due to situations in which the color of the pupil region, in dark eyes, was similar to the color above the pupil region. Another reason was that the shadow influence due to the close proximity of the patient’s body parts relative to the eye position in relation to the light source in the room. The methodology was tested in (1)

Conclusion

This work presented a new method for the detection and diagnosis of strabismus from digital videos through the cover test. This method achieved a strabismus diagnosis with an accuracy of 87% while presenting an average error of 2.57Δ, of the measurements of deviation magnitude, within the error tolerance. With these promising results, we have demonstrated the feasibility of using computational resources based on image processing techniques to achieve this goal. This work will motivate new

Acknowledgments

Our research group acknowledges financial support from FAPEMA (GrantNumber: 007068/2014), CNPQ (GrantNumber: 552108/2011-1) and TECGRAF/PUC-RIO.

References (41)

S.M. Archer et al.
Social and emotional impact of strabismus surgery on quality of life in children
J. Am. Assoc. Pediatr. Ophthalmol. Strabismus
(2005)
E.A. Paysse et al.
Age of the emergence of negative attitudes toward strabismus
J. Am. Assoc. Pediatr. Ophthalmol. Strabismus
(2001)
H.B. McBain et al.
The impact of strabismus on quality of life in adults with and without diplopia: a systematic review
Surv. Ophthalmol.
(2014)
N.A. Barnard et al.
A quantitative analysis of eye movements during the cover testa preliminary report
Ophthalmic Physiol. Optics
(1995)
B.B. Rainey et al.
Reliability of and comparisons among three variations of the alternating cover test
Ophthalmic Physiol. Optics
(1998)
E.M. Helveston et al.
Telemedicine: strabismus e-consultation
J. Am. Assoc. Pediatric Ophthalmol. Strabismus
(2001)
H.A. Johns et al.
The effect of strabismus on a young child’s selection of a playmate
Ophthalmic Physiol. Optics
(2005)
A. Frandson
Occurence of squint. a clinical-statistical study on the prevalence of squint and associated signs in different groups and ages of the danish population
Acta Ophthalmol.
(1960)
P.A. Graham
Epidemiology of strabismus
Br. J. Ophthalmol.
(1974)
P. Mruthyunjaya et al.
Subjective and objective outcomes of strabismus surgery in children
J. Pediatr. Ophthalmol. Strabismus
(1996)

A. Bernfeld

Psychological repercussions of strabismus in children

J. Fr. Ophtalmol.

(1981)

D. Lilakova et al.

Effect of strabismus on the quality of life in adults

Ceska a slovenska oftalmologie: casopis Ceske oftalmologicke spolecnosti a Slovenske oftalmologicke spolecnosti

(2003)

J.R. Anderson

The development of selfrecognition: a review

Dev. Psychobiol.

(1984)

M. Kothari et al.

Comparison of psychosocial and emotional consequences of childhood strabismus on the families from rural and urban india

Indian J. Ophthalmol.

(2004)

K. Wright et al.

Pediatric ophthalmology and strabismus

(2013)

E. Peli et al.

Dynamics of cover test eye movements

Optometry Vision Sci.

(1983)

M.W. Quick et al.

A photographic technique for measuring horizontal and vertical eye alignment throughout the field of gaze

Invest. Ophthalmol. Vis. Sci.

(1992)

I. Subharngkasen

Successful amblyopia therapy by using synoptophore.

J. Med. Assoc. Thailand

(2003)

A. Chandna et al.

Pattern recognition of vertical strabismus using an artificial neural network (strabnet).

Strabismus

(2009)

T.Y. Kim et al.

A new software for quantitative measurement of strabismus based on digital image

J. Korea Multimedia Soc.

(2012)

Cited by (29)

A new digitised screen test for strabismus measurement
2023, Zeitschrift fur Medizinische Physik
Citation Excerpt :
In clinical practice, these traditional examination methods often require time and longstanding practical experience on the part of the examiner. Automated auxiliary systems in the measurement and diagnosis of strabismus can, therefore, contribute to improvement [8–14]. Leitritz et al. [15] presented a digital variant of the Harms tangent screen test with a projection to the wall by a beamer, a pointer for the patient (computer mouse) and two-colour glasses with an integrated forehead projector as a fixation aid for head orientation.
Our study presents a digitised tangent screen test for ocular motility analysis according to the Harms and Hess tests (measurement of the squint angle in all fields of vision). This test uses an image beamer to display the tangent screen, a position sensor to measure the patient’s head orientation, and a distance sensor to measure the fixation distance. Digital measurement of head orientation allows for a test procedure that eliminates the conventional requirement for a light pointer in the patient’s hand. Thus, the digital screen test is presented, and the uncertainty of the measurement system is evaluated.
A mathematical relationship was given between the measured squint angles, as well as the angle of diagnostic gaze direction, and the influence quantities on their measurement uncertainty. The individual uncertainties resulted from deviations in the measured values by the position and distance sensors, the calibration of the projection image of the beamer in length units, and the finite image resolution of the beamer. The individual standard uncertainties of the influence quantities were determined. The combined standard measurement uncertainties of the squint and gaze direction angles were given based on the model equation of the error propagation law at the tangent table according to Harms at a test distance of 2.5 m. The patient’s uncertainty contribution to the mobility analysis was not considered.
The combined standard uncertainty of the measurement system (coverage factor k = 2 for 95% confidence level) for the squint angle is ≤ 0.43° for the angle of diagnostic gaze direction ≤ 3.13° at the test distance of 2.5 m. The individual standard uncertainties of the influence quantities on the angles are (k = 1): 1.55°/1.01° (horizontal/vertical angle of the position sensor), 0.19° (distance sensor), 0.06° (calibration of the projection image of the beamer), and 0.02° (image resolution of the beamer). The maximum valid test distance of the digital screen test is 3.8 m.
The digital screen test is compact and can be used at different locations. Compared to the traditional test, the time required for examination via the digitised test is less; additionally, its documentation is simplified. The measurement uncertainty of the diagnostic gaze direction angle is dominated by the sensor drift of the position sensor in the horizontal direction (yaw angle) and is due to the sensor technology. However, this drift error does not affect the squint angle measurement result nor its measurement uncertainty because the measurement principle used here is based on the congruence between the position cross and the fixation object and the confusion principle and compensates for the drift error. The measurement uncertainties of the determined measurement system are the lower limits of the uncertainties in the clinical use of the digital screen test if there are no effects due to significant patient deviations.
Computational method for aid in the diagnosis of sixth optic nerve palsy through digital videos
2022, Computers in Biology and Medicine
Citation Excerpt :
Finally, Section 6 presents the conclusions and suggestions for future work. Several works have used computational methods to automatically pre-diagnose ocular diseases or to assist the ophthalmologist in their diagnosis [4–7,9,11,13,21]. However, when it comes to sixth optic nerve palsy (SONP), computational methods aimed directly at detecting or diagnosing this palsy using digital images or videos have not been found until writing this paper using traditional search means.
The sixth cranial nerve, also known as the abducens nerve, is responsible for controlling the movements of the lateral rectus muscle. Palsies on the sixth nerve prevent some muscles that control eye movements from proper functioning, causing headaches, migraines, blurred vision, vertigo, and double vision. Hence, such palsy should be diagnosed in the early stages to treat it without leaving any sequela. The usual methods for diagnosing the sixth nerve palsy are invasive or depend on expensive equipment, and computer-based methods designed specifically to diagnose the aforementioned palsy were not found until the publication of this work. Therefore, a low-cost, non-invasive method can support or guide the ophthalmologist’s diagnosis. In this context, this work presents a computational methodology to aid in diagnosing the sixth nerve palsy using videos to assist ophthalmologists in the diagnostic process, serving as a second opinion. The proposed method uses convolutional neural networks and image processing techniques to track both eyes’ movement trajectory during the video. With this trajectory, it is possible to calculate the average velocity (AV) in which each eye moves. Since it is known that paretic eyes move slower than healthy eyes, comparing the AV of both eyes can determine if the eye is healthy or paretic. The results obtained with the proposed method showed that paretic eyes move at least 19.65% slower than healthy ones. This threshold, along with the AV of the movement of the eyes, can help ophthalmologists in their analysis. The proposed method reached 92.64% accuracy in diagnosing the sixth optic nerve palsy (SONP), with a $K a p p a$ index of 0.925, which highlights the reliability of the results and gives favorable perspectives for further clinical application.
Automated Diagnosis and Measurement of Strabismus in Children
2020, American Journal of Ophthalmology
Citation Excerpt :
Furthermore, using a nonaccommodative (a spot of light) target and asymmetric kappa angles increased the susceptibility to errors. Another recent pilot trial using a digital video analysis methodology that interprets shadowing related to room light, used for detecting and diagnosing exotropia, was compromised by room light conditions unsuitable for dark eyes.19 Furthermore, the validity of strabismus testing using video goggles was recently assessed in comparison with the Hess screen test.
Manual measurements of strabismus are subjective, time consuming, difficult to perform in babies, toddlers, and young children, and rely on the examiner's skill and experience. An automated system, based on eye tracking and dedicated full occlusion glasses, was developed to provide a fast, objective, and easy-to-use alternative to the manual measurements of strabismus. This study tested the efficacy of the system in determining the presence of strabismus in children, as well as its type and the amount of deviation, in addition to differentiating between phorias and tropias.
A prospective, masked, inter-rater reliability study.
A prospective, masked, cross-sectional study included 69 children, 3-15 years of age. A cover-uncover test and a prism alternating cover test (PACT) for the primary gaze, at a distance of 50 cm, were performed by 2 independent, masked examiners and by the automated system.
A high correlation was found between the automated and the manual test results (R = 0.9 and P < 0.001 for the horizontal deviation, and R = 0.91 and P < 0.001 for the vertical deviations, with 100% correct identification of the type of deviation). The average automated test duration was 46 seconds. The Bland-Altman plot, used to compare the 2 measurement methods, showed a mean value of −2.9 prism diopters (PD) and a half-width of the 95% limit of agreement of ±11.4 PD.
The automated system provides precise detection and measurements of ocular misalignment and differentiated between phorias and tropias. It can be used in participants from the ages of 3 years old and up.
Virtual reality-based measurement of ocular deviation in strabismus
2020, Computer Methods and Programs in Biomedicine
Citation Excerpt :
However, HT requires the Hirschberg ratio and angle Kappa that are subject-independent, which significantly affects the accuracy. Techniques based on the alternating cover test (ACT) are proposed in [16–18], where the ocular deviation angle is generated from the eye movement while alternating the cover. Chin rest must be used for resisting the relative position change between a patient and a target throughout the experiment, however, small errors are avoidable.
Strabismus is an eye movement disorder in which shows the abnormal ocular deviation. Cover tests have mainly been used in the clinical diagnosis of strabismus for treatment. However, the whole process depends on the doctor's level of experience, which could be subjected to several factors. In this study, an automated technique for measurement of ocular deviation using a virtual reality (VR) device is developed.
A VR display system in which the screens that have the fixation target are changed alternately between on and off stages is used to simulate the normal strabismus diagnosis steps. Patients watch special-designed 3D scenes, and their eye motions are recorded by two infrared (IR) cameras. An image-processing-based pupil tracking technique is then applied to track their eye movement. After recording eye motion, two strategies for strabismus angle estimation are implemented: direct measurement and stepwise approximation. The direct measurement converts the eye movement to a strabismus angle after considering the eyeball diameter, while the stepwise approximation measures the ocular deviation through the feedback calibration process.
Experiments are carried out with various strabismus patients. The results are compared to those of their doctors’ measurement, which shows good agreement.
The results clearly indicate that these techniques could identify ocular deviation with high accuracy and efficiency. The proposed system can be applied in small space and has high tolerance for the unexpected head movements compared with other camera-based system.
Detection of white matter lesion regions in MRI using SLIC0 and convolutional neural network
2018, Computer Methods and Programs in Biomedicine
Citation Excerpt :
In order to support specialists in the identification of anomalies in human body, many researches have invested substantially in the development and usage of image processing and pattern recognition techniques. These techniques have been combined to develop Computer-Aided Detection (CAD) and Computer-Aided Diagnostic (CADx) systems and applied in several medical images types, such as CT and MRI, to provide support in the analysis of several types of disease [66–71]. These systems provide additional information, increase analysis efficiency and reduce workload of the specialists.
Background and Objective: White matter lesions are non-static brain lesions that have a prevalence rate up to 98% in the elderly population. Because they may be associated with several brain diseases, it is important that they are detected as soon as possible. Magnetic Resonance Imaging (MRI) provides three-dimensional data with the possibility to detect and emphasize contrast differences in soft tissues, providing rich information about the human soft tissue anatomy. However, the amount of data provided for these images is far too much for manual analysis/interpretation, representing a difficult and time-consuming task for specialists. This work presents a computational methodology capable of detecting regions of white matter lesions of the brain in MRI of FLAIR modality. The techniques highlighted in this methodology are SLIC0 clustering for candidate segmentation and convolutional neural networks for candidate classification. Methods: The methodology proposed here consists of four steps: (1) images acquisition, (2) images preprocessing, (3) candidates segmentation and (4) candidates classification. Results: The methodology was applied on 91 magnetic resonance images provided by DASA, and achieved an accuracy of 98.73%, specificity of 98.77% and sensitivity of 78.79% with 0.005 of false positives, without any false positives reduction technique, in detection of white matter lesion regions. Conclusions: It is demonstrated the feasibility of the analysis of brain MRI using SLIC0 and convolutional neural network techniques to achieve success in detection of white matter lesions regions.
Usefulness Assessment of Automated Strabismus Angle Measurements Using Innovative Strabiscan Device
2024, Journal of Clinical Medicine

View all citing articles on Scopus

View full text

Automatic diagnosis of strabismus in digital videos through cover test

Highlights

Abstract

Introduction

Section snippets

Related work

Proposed methodology

Results

Discussion

Conclusion

Acknowledgments

J. Am. Assoc. Pediatr. Ophthalmol. Strabismus

J. Am. Assoc. Pediatr. Ophthalmol. Strabismus

Surv. Ophthalmol.

Ophthalmic Physiol. Optics

Ophthalmic Physiol. Optics

J. Am. Assoc. Pediatric Ophthalmol. Strabismus

The effect of strabismus on a young child’s selection of a playmate

Ophthalmic Physiol. Optics

Occurence of squint. a clinical-statistical study on the prevalence of squint and associated signs in different groups and ages of the danish population

Acta Ophthalmol.

Epidemiology of strabismus

Br. J. Ophthalmol.

Subjective and objective outcomes of strabismus surgery in children

J. Pediatr. Ophthalmol. Strabismus

Psychological repercussions of strabismus in children

J. Fr. Ophtalmol.

Effect of strabismus on the quality of life in adults

Ceska a slovenska oftalmologie: casopis Ceske oftalmologicke spolecnosti a Slovenske oftalmologicke spolecnosti

The development of selfrecognition: a review

Dev. Psychobiol.

Comparison of psychosocial and emotional consequences of childhood strabismus on the families from rural and urban india

Indian J. Ophthalmol.

Pediatric ophthalmology and strabismus

Dynamics of cover test eye movements

Optometry Vision Sci.

A photographic technique for measuring horizontal and vertical eye alignment throughout the field of gaze

Invest. Ophthalmol. Vis. Sci.

Successful amblyopia therapy by using synoptophore.

J. Med. Assoc. Thailand

Pattern recognition of vertical strabismus using an artificial neural network (strabnet).

Strabismus

A new software for quantitative measurement of strabismus based on digital image

J. Korea Multimedia Soc.