Vision-Based Diagnosis and Treatment

Abstract

Devices and methods are provided herein for aiding in the diagnosis of and response to conditions ranging from unpleasant to disabling. Images are presented to viewers and eye-tracking of their eyes during that process along with software analysis non-subjectively identify and quantify complex visual signatures formerly hidden in the responses of the eyes. This allows caregivers, when presenting these images to people to be evaluated for a condition, to identify, again with eye-tracking and software analysis, those signatures indicative of a presence or an absence of the condition. An associated remediative response provides treatment applicable to a real-time environment.

What is claimed is: 
     
         1 . A device for identifying characteristics of a condition comprising:
 a plurality of images;   a display device to display said plurality of images;   viewer data that indicates, for a plurality of viewers of said plurality of images, at least their status regarding the condition;   an eye-tracking assembly; and   a processor assembly configured to determine, from said viewer data and where said plurality of viewers look on those images according to said eye-tracking assembly, identified locations where viewers having essentially the same said status tend to look.   
     
     
         2 . The device of  claim 1 , wherein
 if a subsequent viewer of at least one of those images looks at one of said identified locations it can be used as at least an indication of the status of that viewer regarding at least if that subsequent viewer has the condition.   
     
     
         3 . The device of  claim 1 , wherein
 said plurality of images are images in video form; and   said identified locations determined by said processor assembly identify at least one of   1) at least one point in one of said plurality of images or   2) an area in one of said plurality of images.   
     
     
         4 . The device of  claim 1 , further comprising
 a data storage device to store results for future use; wherein   the evaluation of a subsequent viewer viewing at least some of said plurality of images can be assisted by identifying at least if that subsequent viewer looked at one of the said identified locations indicated by this stored data.   
     
     
         5 . The device of  claim 1 , wherein
 wherein said viewer data at least contains data on said status indicative of at least one of:
 1) which of said plurality of viewers have said condition. 
 2) which of said plurality of viewers do not have said condition or 
 3) to what degree said plurality of viewers have said condition. 
   
     
     
         6 . The device of  claim 5 , wherein
 at least one of said identified locations determined by said processor assembly are given at least one scoring value by said processor assembly based on at least one of   1) how many of said plurality of viewers having a similar said status looked there,   2) how long they looked there,   3) different degrees of said condition indicated in said viewer data for those of said plurality of viewers who looked there,   4) different degrees of the absence of said condition indicated in said viewer data for those of said plurality of viewers who looked there or   5) the size of the identified locations; whereby   the magnitude of the significance of a person looking at one of said identified locations is more precisely establishable.   
     
     
         7 . The device of  claim 1 , wherein
 said processor assembly estimates, for at least one of said identified locations a value indicative of, at least one of   1) the likelihood that a subsequent viewer looking there has said condition or   2) the degree to which a subsequent viewer looking there has said condition,   based on at least said viewer data and where those with a known status with respect to said condition looked.   
     
     
         8 . The device of  claim 1 , wherein
 to determine said identified locations said processor assembly uses a process at least partially analogous to cluster analysis to identify where at least some of said plurality of viewers having at least a partially shared status with respect to said condition often look.   
     
     
         9 . The device of  claim 1 , wherein
 said processor assembly identifies at least one of said identified locations at least partially by identifying which of a variety of possible locations would most closely return, in a calculation to identify the statuses of said plurality of viewers based at least in part on what locations they looked at, the same statuses found in said viewer data.   
     
     
         10 . The device of  claim 1 , wherein
 said processor assembly identifies at least one of said identified locations and parameters associated with it at least partially by identifying which of a variety of possible locations and parameters would most closely return, in a calculation to identify the statuses of said plurality of viewers based at least in part on what locations they looked at, the same statuses found in said viewer data.   
     
     
         11 . The device of  claim 1 , wherein
 said processor assembly determines said identified locations and scoring values for them by calculating at least which scoring values and locations are usable to identify, based on where said plurality of viewers looked, statuses, regarding said condition for those viewers, that best match the ones indicated by said viewer data; whereby   if a subsequent viewer of any of said plurality of images looks at one of the thus-identified said identified locations, that is at least an indication of a status of said condition and the scoring values are useful in estimating a magnitude of the significance of that indication.   
     
     
         12 . A method for determining diagnostic indicators of a condition comprising the steps of:
 presenting a scene to a viewer whose status regarding the condition is available;   identifying and noting where in that scene the viewer looks;   continuing with another scene, if there are more scenes to be shown, at the step of presenting a scene to a viewer whose status regarding the condition is available;   continuing with the next viewer, if there are additional viewers to be included, at the step of presenting a scene to a viewer whose status regarding the condition is available;   calculating, for at least some of the scenes, diagnostic areas where those having a similar status with respect to the condition tend to look.   
     
     
         13 . The method of  claim 12  wherein
 the location of and parameters for said diagnostic areas that, in a calculation using them to identify the status of a viewer based on where that viewer looks, would result in a calculated status for those previous viewers that is closest to the available said statuses for them is accomplished as part of the step of 
 calculating, for at least some of the scenes, diagnostic areas where those having a similar status with respect to the condition tend to look. 
 
     
     
         14 . The method of  claim 13 , wherein
 said parameters for said diagnostic areas are at least indicative of at least one of   1) the scope of the diagnostic area,   2) the scoring significance of a viewer looking at the diagnostic area,   3) the scoring significance of a viewer looking at a location beyond the diagnostic area,   4) the shape of the diagnostic area or   5) weighting values for use in increasing accuracy.   
     
     
         15 . The method of  claim 12  further comprising the steps of
 presenting one of said scenes that has at least said diagnostic areas identified to a new viewer; 
 identifying where in the scene this new viewer looks; 
 determining, based at least on where this new viewer looks with respect to said diagnostic areas, at least an indication of the new viewer's status with respect to said condition. 
 
     
     
         16 . The method of  claim 15  further comprising the steps of
 alerting said new viewer, responsive to said indication of the new viewer's status, with the form of that alerting being at least one of 
 1. affecting the visibility of the displayed said scene, 
 2. graphically indicating a preferred location to look at or 
 3. graphically indicating a preferred direction to look towards; and 
 continuing, if there are more scenes to be displayed, at the step of presenting one of said scenes that has at least said diagnostic areas identified to a new viewer. 
 
     
     
         17 . The method of  claim 15  further comprising the steps of
 storing at least data indicative of said status; 
 continuing with another said scene, if more scenes are desirable, at the step of presenting one of said scenes that has at least said diagnostic areas identified to a new viewer; and 
 calculating, based on determined status data, a diagnostic measure indicative of one of 
 1) a measure of the presence of said condition in this new viewer or 
 2) a measure of the absence of said condition in this new viewer. 
 
     
     
         18 . A device for identifying characteristics of a condition comprising:
 at least one image;   a display device to display said at least one image;   viewer data for a plurality of viewers of said at least one image indicating at least which have said condition;   an eye-tracking assembly;   a processor to at least identify, from the eye-tracking assembly and viewer data, portions of said at least one image where those in said plurality of viewers who have the condition are likely to look.   
     
     
         19 . The device of  claim 18 , wherein
 if a subsequent viewer of said at least one image looks at one of said portions it can be used as at least one indication of at least if that viewer has said condition.   
     
     
         20 . The device of  claim 18 , wherein
 said at least one image is in video form.   
     
     
         21 . The device of  claim 18 , wherein
 said viewer data at least contains data indicative of at least one of   1) which of said plurality of viewers have said condition,   2) which of said plurality of viewers do not have said condition or   3) to what degree at least one of said plurality of viewers has said condition; and   said processor also identifies portions of said at least one image where those in said plurality of viewers who do not have said condition are likely to look.   
     
     
         22 . The device of  claim 21 , wherein
 said processor calculates scoring parameters, associated with said portions of said at least one image, responsive to at least one of   1) how many of said plurality of viewers having said condition looked there,   2) how many of said plurality of viewers not having said condition looked there,   3) to what degree those of said plurality of viewers who looked there had said condition,   4) how long viewers having said condition looked there,   5) how long viewers not having said condition looked there or   6) the size of the portion of said at least one image.   
     
     
         23 . The device of  claim 22 , wherein
 said scoring parameters for one of said portions of said at least one image will, if a subsequent viewer looks there, provide at least an indication of the status of that subsequent viewer regarding any presence of said condition.   
     
     
         24 . The device of  claim 18 , wherein
 to identify one of said portions of said at least one image in said plurality of images, said processor locates statistical densities in portions of an image where those in said plurality of viewers having a similar status regarding said condition tend to look; whereby   this process identifies portions of that image indicative of a status of said condition when a viewer looks there.   
     
     
         25 . The device of  claim 18 , wherein
 said processor determines said portions of said at least one image and scoring parameters for them by calculating at least which of a variety of possible scoring parameters and portions would, in a calculation to identify statuses based on where said plurality of viewers looked, result in calculated statuses that best match those indicated by said viewer data.