TY - GEN
T1 - MobileVision
T2 - 5th Conference on Wireless Health, WH 2014
AU - Samaniego, Adam
AU - Boominathan, Vivek
AU - Sabharwal, Ashutosh
AU - Veeraraghavan, Ashok
N1 - Publisher Copyright:
Copyright © 2014 ACM.
PY - 2014/10/29
Y1 - 2014/10/29
N2 - We present mobileVision - A portable, robust, smartphonebased ophthalmoscopy prototype intended to reduce the barriers to ocular pathology screening in developing and underserved regions. In contrast to currently available retinal imaging solutions, mobileVision provides the portability of a handheld ophthalmoscope without sacrificing retinal field-of-view or resolution. Through tight integration with a smartphone and ergonomic design, we demonstrate novel features for such a small form factor, including: Automatic compensation for patient refractive error, voice-activated multi-shot retinal image acquisition without pupil dilation (non-mydriatic), and touch-gesture based control of patient fixation and accommodation. We further demonstrate a computational lucky imaging and retinal stitching pipeline which not only increases overall retinal field-of-view, but also makes the system robust to patient saccades, blinks, device jitter, and imaging artifacts such as noise or unintended scattering from ocular surfaces. We estimate through mock eye tests that the mobileVision prototype is capable of imaging the retina with 23.5μm of retinal resolution for patients with between -6D to +13D refractive error, and we image over ±45° of retina during an in vivo trial.
AB - We present mobileVision - A portable, robust, smartphonebased ophthalmoscopy prototype intended to reduce the barriers to ocular pathology screening in developing and underserved regions. In contrast to currently available retinal imaging solutions, mobileVision provides the portability of a handheld ophthalmoscope without sacrificing retinal field-of-view or resolution. Through tight integration with a smartphone and ergonomic design, we demonstrate novel features for such a small form factor, including: Automatic compensation for patient refractive error, voice-activated multi-shot retinal image acquisition without pupil dilation (non-mydriatic), and touch-gesture based control of patient fixation and accommodation. We further demonstrate a computational lucky imaging and retinal stitching pipeline which not only increases overall retinal field-of-view, but also makes the system robust to patient saccades, blinks, device jitter, and imaging artifacts such as noise or unintended scattering from ocular surfaces. We estimate through mock eye tests that the mobileVision prototype is capable of imaging the retina with 23.5μm of retinal resolution for patients with between -6D to +13D refractive error, and we image over ±45° of retina during an in vivo trial.
KW - Algorithms
KW - Design
UR - http://www.scopus.com/inward/record.url?scp=84919372210&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=84919372210&partnerID=8YFLogxK
U2 - 10.1145/2668883.2668886
DO - 10.1145/2668883.2668886
M3 - Conference contribution
AN - SCOPUS:84919372210
T3 - Proceedings - Wireless Health 2014, WH 2014
BT - Proceedings - Wireless Health 2014, WH 2014
PB - Association for Computing Machinery
Y2 - 29 October 2014 through 31 October 2014
ER -