Patient-attentive sequential strategy for perimetry-based visual field acquisition

•We formulate perimetry testing problem as a sequential experimental design problem.•Our strategy, PASS, selects locations based on the patient’s previous answers and reconstructs visual fields from sparse observations.•PASS is patient-specific and non-greedy.•Experiments show that PASS provides bet...

Full description

Saved in:
Bibliographic Details
Published in:Medical image analysis 2019-05, Vol.54, p.179-192
Main Authors: Kucur, Şerife Seda, Márquez-Neila, Pablo, Abegg, Mathias, Sznitman, Raphael
Format: Article
Language:English
Subjects:
Exhausting
Field of view
Image reconstruction
Neural network
Perimetry strategy
Reinforcement learning
Sequential experimental design
Visual field
Visual fields
Visual observation
Visual stimuli
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:•We formulate perimetry testing problem as a sequential experimental design problem.•Our strategy, PASS, selects locations based on the patient’s previous answers and reconstructs visual fields from sparse observations.•PASS is patient-specific and non-greedy.•Experiments show that PASS provides better accuracy-speed tradeoff than the state-of-the-art methods. [Display omitted] PASS, a patient-specific and non-greedy strategy, adaptively selects what locations to query based on the patient’s answers to previous queries, and the locations are jointly selected to maximize the quality of the final visual field (VF) reconstruction. Each time a new location is selected, the selected location is observed and used along with the previous observations to reconstruct the VF. The reconstructed VF and the history of observed locations are used to select the next location to observe. This iterative strategy is stopped when the intended number of locations are observed. The procedure is shown to attend relevant VF locations and to lead to accurate and fast VF acquisitons. Perimetry is a non-invasive clinical psychometric examination used for diagnosing ophthalmic and neurological conditions. At its core, perimetry relies on a subject pressing a button whenever they see a visual stimulus within their field of view. This sequential process then yields a 2D visual field image that is critical for clinical use. Perimetry is painfully slow however, with examinations lasting 7–8 minutes per eye. Maintaining high levels of concentration during that time is exhausting for the patient and negatively affects the acquired visual field. We introduce PASS, a novel perimetry testing strategy, based on reinforcement learning, that requires fewer locations in order to effectively estimate 2D visual fields. PASS uses a selection policy that determines what locations should be tested in order to reconstruct the complete visual field as accurately as possible, and then separately reconstructs the visual field from sparse observations. Furthermore, PASS is patient-specific and non-greedy. It adaptively selects what locations to query based on the patient’s answers to previous queries, and the locations are jointly selected to maximize the quality of the final reconstruction. In our experiments, we show that PASS outperforms state-of-the-art methods, leading to more accurate reconstructions while reducing between 30% and 70% the duration of the patient examination.
ISSN:1361-8415
1361-8423
DOI:10.1016/j.media.2019.03.002