Photorefraction ocular screening device and methods
View Patent ↗A photorefraction ocular screening device for assessing vision and corresponding disorders associated with the human ocular system is provided. More specifically, the present invention provides for a photorefraction ocular screening device employing advanced methods of pupil detection and refractive error analysis. The photorefraction ocular screening device is comprised of an LED arrangement configured with a plurality of irradiation sources serving as visual stimuli, wherein the visual stimuli may be presented in varying illumination patterns to the pupils of an examinee for expanding the range of ocular responses that can be used to determine refractive error.
1. A computer-implemented method of acquiring pupils, said method comprising the steps of:
receiving a full frame image captured by an image capture component;
decimating said full frame image to produce a decimated image;
applying a pupil filter to said decimated image to identify pupil pixels;
acquiring a pupil candidate from said filtered decimated image;
identifying a corresponding region of interest of said pupil candidate in said full frame image;
binarizing said region of interest to identify pixels representative of pupil edge; and
applying a best-fit circle to said pixels representative of pupil edge.
2. The method of claim 1 , wherein said decimated image is smaller in area than said full frame image.
3. The method of claim 2 , wherein said decimated image is 1/16.sup.th the area of said full frame image.
4. The method of claim 1 , wherein said pupil filter is applied to said decimated image to aid in determining pupil and non-pupil regions.
5. The method of claim 4 , wherein said pupil filter is optimized for speed by considering a limited-number of symmetrically-spaced surrounding pixels in said decimated image.
6. The method of claim 1 , wherein a bounding rectangle is defined by a left-most column, a right-most column, a top-most row and a bottom-most row of said pupil candidate.
7. The method of claim 6 , wherein said pupil candidate is rejected if said bounding rectangle comprises a width dimension or a height dimension that is less than two pixels in size.
8. The method of claim 6 , wherein a center of said bounding rectangle is adjusted to match an equivalent position in said full frame image to establish said region of interest.
9. The method of claim 1 , wherein binarizing said region of interest further comprises combining results yielded by a first edge-detection approach and a second edge-detection approach conducted on said region of interest in said full frame image.
10. The method of claim 9 , wherein said first edge-detection approach employs an intensity histogram to identify brightness and darkness intensity levels in said region of interest.
11. The method of claim 9 , wherein said second edge-detection approach employs a pseudo-gradient calculation using adjacent pixels to identify a rate of change of brightness and darkness intensity levels across said region of interest.
12. The method of claim 9 , wherein combining said first edge-detection approach and said second edge-detection approach only retains edge pixels common in results yielded by both of said approaches.
13. The method of claim 1 , wherein said best-fit circle is computed using a least squares calculation.
14. A system employing computer processing logic for acquiring pupils, comprising:
a processor component coupled to at least one memory component, said processor component configured to:
receive a full frame image captured by an image capture component;
decimate said full frame image;
apply a pupil filter to said decimated image to identify pupil pixels;
acquire a pupil candidate from said filtered decimated image;
identify a corresponding region of interest of said pupil candidate in said full frame image;
binarize said region of interest to identify pixels representative of pupil edge; and
apply a best-fit circle to said pixels representative of pupil edge.
15. A non-transitory computer-readable storage medium programmed to include instructions that, when executed by a processor, cause the processor to perform a method of acquiring pupils, said method comprising the steps of:
receiving a full frame image captured by an image capture component;
decimating said full frame image;
applying a pupil filter to said decimated image to identify pupil pixels;
acquiring a pupil candidate from said filtered decimated image; identifying a corresponding region of interest of said pupil candidate in said full frame image;
binarizing said region of interest to identify pixels representative of pupil edge; and
applying a best-fit circle to said pixels representative of pupil edge.