Prediction of visual acuity from wavefront aberrations
AssigneeThe United States of America as Represented by the Administrator of the National Aeronautics & Space Administration (NASA)
InventorsAndrew B. Watson, Albert J. Ahumada
Life of this patent
Estimating the term…
Abstract
A method for generating a visual acuity metric, based on wavefront aberrations (WFAs), associated with a test subject and representing classes of imperfections, such as defocus, astigmatism, coma and spherical aberrations, of the subject's visual system. The metric allows choices of different image template, can predict acuity for different target probabilities, can incorporate different and possibly subject-specific neural transfer functions, can predict acuity for different subject templates, and incorporates a model of the optotype identification task.
Claims6
A method of predicting visual acuity of an imaging system, the method comprising:
- providing a set of K optotypes, numbered k=1,. . . , K (K≧2) of a specified size, to be used to establish a reference set of symbols to estimate visual acuity, and providing a description or image Tk(x,y) of each optotype, dependent upon location coordinates (x,y) associated with the optptype image produced by the imaging system;
- constructing a Wavefront Aberration Image WAI(x,y) that manifests selected image aberrations associated with the imaging system;
- computing a generalized pupil image GPI(x,y), with non-zero values confined to within a pupil aperture associated with the imaging system, that manifests the selected image aberrations and that is dependent upon at least one wavelength λ of light with which the image is viewed;
- computing a Point Spread Function PSF(x,y) that is an absolute value squared of a Discrete Fourier Transform (DFT) of the image GPI(x,y);
- computing a Normalized Point Spread Function NPSF(x,y), proportional to the Point Spread Function PSF (x,y), whose norm is 1:
- computing an Optical Transfer Function OTF(u,v), expressed as a function of spatial frequency indice (u,v) in a Fourier transform plane, that is a Discrete Fourier Transform of the Normalized Point Spread Function NPSF (x,y), multiplied by a value h of original image size and multiplied by an image magnification index m;
- generating a Radial Neural Transfer Function RNTF(u,v), where RNTF(u,v) satisfies the following conditions:
- (i) RNTF(u,v) is a continuous, non-negative function of a spatial frequency variable f=[u2+v2]1/2;
- (ii) RNTF(u,v) has at least one value, f=f(max), for which NRTF(u,v) is a maximum;
- (iii) for 0<f<f(max), RNTF(u,v) is monotonically increasing in f;
- (iv) for f>f(max), and RNTF(u,v) is monotonically decreasing toward 0;
- computing a Neural Transfer Function NTF(u,v) as a product of the Radial Neural Transfer Function RNTF (u,v) and an Oblique Effect Filter function OER(u,v) that compensates for viewing angle of the-original optotype image;
- generating a Total Transfer Function TTF(u,v), defined as a product of NTF(u,v) and the Optical Transfer Function OTF(u,v);
- generating a Proportion Correct index PC(j), which presents a probability associated with a correct optotype that would be identififed by a subject having the Wavefront Aberration Image WAI(x,y) for at least one optotype(j).
Dependencies are inferred from the claim text.