Quantitative disorder enhanced augmentative/alternative communication device and process
View Patent ↗A method of interpreting human-drawn images includes modifying a human-drawn image. A numerical value corresponding to an order parameter squared (S 2 ) is extracted from the modified image. An artificial intelligence (AI) program characterizes the human-drawn image utilizing the human-drawn image and the numerical value of the order parameter. The disclosure further includes systems, computer readable media, programs capable of the same.
1 . A computer-implemented method of recognizing the content of a human-drawn image, the method comprising:
utilizing a computer to form a modified image from a human-drawn image by replicating a physical image formation process;
utilizing a computer to extract a numerical value corresponding to an order parameter squared (S 2 ) from the modified image, wherein the modified image includes light regions and dark regions, and wherein S 2 comprises a numerical value quantifying a degree of order present in the modified image, and wherein the extracted numerical value comprises a ratio of an area of the light regions to a total area that is equal to the sum of: 1) an area of the light regions, and 2) of an area of the dark regions;
causing an artificial intelligence (AI) program to characterize the human-drawn image utilizing: 1) the human-drawn image, and 2) the numerical value of the order parameter extracted from the modified image formed from the human-drawn image; and:
causing the computer to output at least one of an image and text that identifies the human-drawn image.
2 . The method of claim 1 , further including:
training the AI program by causing the AI program to characterize human-drawn images for a plurality of non-identical human-drawn images using 1) each of the plurality of non-identical human-drawn images, and 2) the numerical value of the order parameter extracted from the modified image formed from each of the plurality of non-identical human-drawn images.
3 . The method of claim 2 , wherein:
the plurality of non-identical human-drawn images are formed by a human utilizing an input device that allows a user to manually form the human-drawn image data.
4 . The method of claim 1 , wherein:
the computer comprises a portable device having a touch screen;
the human-drawn image data includes at least one image drawn by a human using the touch screen.
5 . The method of claim 4 , wherein:
the portable device is selected from the group consisting of smart phones and tablet computers.
6 . The method of claim 1 , wherein:
the human-drawn image comprises a symbolic drawing of an object;
the AI program characterizes the symbolic drawing by outputting a word describing the object.
7 . The method of claim 1 , wherein:
the human-drawn image comprises text;
the AI program characterizes the symbolic drawing by outputting a word describing the text.
8 . The method of claim 1 , wherein:
causing the AI program to characterize the human-drawn image includes
supplying the AI with 1) the human-drawn image, and 2) the numerical value of the order parameter extracted from the modified image formed from the human-drawn image.
9 . The method of claim 1 , wherein:
the computer comprises a draw-to-speech device;
the human-drawn image represents at least one of numbers, letters, words, pictures, or concepts; and including:
causing the draw-to-speech device to generate an audio signal comprising a word corresponding the numbers, letters, words, pictures, or concepts of the human-drawn image.
10 . The method of claim 1 , wherein utilizing a computer to form a modified image from a human-drawn image includes:
utilizing a computer to form a Fourier spectrum by taking a Fourier transform of the human-drawn image that is in the form of digital image data;
utilizing a computer to form an MTF-modified Fourier transform by applying an idealized modulation transfer function (MTF) to the Fourier spectrum, wherein the MTF is constant across all frequencies;
utilizing a computer to form a modified image by taking an inverse Fourier transform of the MTF-modified Fourier transform.
11 . A data processing system for recognizing the content of a human-drawn image comprising means for
forming a modified image from a human-drawn image by replicating a physical image formation process;
extracting a numerical value corresponding to an order parameter squared S 2 from the modified image, wherein the modified image includes light regions and dark regions, and wherein S 2 comprises a numerical value quantifying a degree of order present in the modified image, and wherein the extracted numerical value comprises a ratio of an area of the light regions to a total area that is equal to the sum of: 1) an area of the light regions, and 2) of an area of the dark regions;
using an artificial intelligence (AI) program to characterize the human-drawn image utilizing: 1) the human-drawn image, and 2) the numerical value of the order parameter extracted from the modified image formed from the human-drawn image; and:
outputting at least one of an image and text that identifies the human-drawn image.
12 . The system of claim 11 wherein the system comprises a draw-to-speech device.
13 . The system of claim 11 , wherein said human-drawn image represents at least one of numbers, letters, words, pictures, or concepts; and the draw-to-speech device is capable of generating an audio signal comprising a word corresponding to the numbers, letters, words, pictures, or concepts of the human-drawn image.
14 . The system of claim 11 , wherein the system comprises a portable device having a touch screen; and
the human-drawn image data includes at least one image drawn by a human using the touch screen.
15 . The system of claim 14 , wherein:
the portable device is selected from the group consisting of smart phones and tablet computers.
16 . The system of claim 11 , wherein forming a modified image from a human-drawn image includes:
utilizing a computer to form a Fourier spectrum by taking a Fourier transform of the human-drawn image that is in the form of digital image data;
utilizing a computer to form an MTF-modified Fourier transform by applying an idealized modulation transfer function (MTF) to the Fourier spectrum, wherein the MTF is constant across all frequencies;
utilizing a computer to form a modified image by taking an inverse Fourier transform of the MTF-modified Fourier transform.