Purchase this article with an account.
H.P. Apple, E.E. Hartmann, J.M. Miller, D.C. Apple; Automated Cover/Uncover Test: Evaluation of Two Prototypes . Invest. Ophthalmol. Vis. Sci. 2003;44(13):4835.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Purpose: The cover/uncover test is the most widely accepted clinical technique for detecting pathologic misalignment of the eyes. This test is often recommended as part of a vision screening protocol for infants and preschool-aged children. In the hands of a skilled practitioner (pediatric optometrist or ophthalmologist) this test provides excellent information for identifying binocular vision problems. However, even a skilled general pediatric practitioner may have difficulty accurately conducting this assessment. A device that automates this assessment in a preschool vision screening context could have substantial utility. Methods: Two prototypes capable of automatically performing various cover test sequences were designed. Electronically controlled liquid crystal shutters silently perform the occlusion sequences and infrared eye-tracking tracks the eyes, which may move in response to the shutter stimulus. The Remote Camera Unit (RCU) version uses a modified video conference camera. The subject wears LCD shutter glasses and looks at real near and far targets. The camera automatically aims, zooms, and focuses into the eyes/glasses region. A digital movie is recorded while the preprogrammed occlusion sequence is performed. The HandHeld Unit (HHU) is similar in shape to a ViewMasterTM stereo viewer. The subject holds it up to his/her face and looks through a glass window at real targets. No virtual targets are involved. This unit employs the LCD shutters and two inexpensive infrared video surveillance cameras. An advantage of the HHU is that it provides a very high-resolution image of each eye. Both units produce an AVI movie, which is postprocessed by a MATLABTM program. Glints and pupil objects are identified using blob analysis methods. Derived image parameters are produced and are analyzed via Kalman filtering and logical rules to produce a final pass/refer recommendation. Deviations are in prism diopters. Results: 4 normal adults and 3 adults with ocular alignment abnormalities were pre-selected. Task compliance was achieved for all subjects with no "movie" retakes required. Screening performance of the automated image analysis was correct in all cases versus the gold standard observer. Conclusions: Early feasibility of both prototypes has been demonstrated. Further experience is now required on preschoolers. It is expected with the preschool population that one of the prototypes may be preferred based on human factors considerations including age.
This PDF is available to Subscribers Only