Image processing and text analysis to determine medical condition

1. A computer-implemented method of identifying a medical condition of a subject, the computer-implemented method comprising:

correlating, by at least one processing device, one or more medical descriptions extracted from textual medical information related to the subject with characteristics of an image of a medical condition of the subject to generate a subject signature, the correlating further comprising:

analyzing the image of the medical condition of the subject to produce a feature vector representing features found in the image,

produce a vector based on the one or more medical descriptions, each element of the vector having a first value representing a respective detectable image feature that is included in the one or more medical descriptions, and each element of the vector having a second value representing a respective detectable image feature that is not included in the one or more medical descriptions, and

generate the subject signature based on the vector and the feature vector;

comparing, by the at least one processing device, the subject signature with a plurality of reference signatures to determine a closest matching reference signature to the subject signature, wherein each reference signature is based on a corresponding second vector produced based on one or more reference medical descriptions extracted from textual medical information associated with a corresponding medical condition and a second feature vector representing features found in an image of the corresponding medical condition;

identifying, by the at least one processing device, the medical condition of the subject based on the corresponding medical condition associated with the determined closest matching reference signature; and

providing information regarding the identified medical condition of the subject.

2. The computer-implemented method of claim 1 , wherein the medical condition includes a disease.

3. The computer-implemented method of claim 1 , further comprising:

generating the plurality of reference signatures, the generating the plurality of reference signatures further comprising:

performing analytics on each of a plurality of textual medical information corresponding to a plurality of medical conditions to extract respective one or more reference medical descriptions of each of the plurality of medical conditions;

analyzing images of the plurality of medical conditions to extract respective image characteristics of each of the plurality of medical conditions; and

correlating the extracted respective one or more reference medical descriptions of each of the plurality of medical conditions with the corresponding extracted image characteristics to generate each of the reference signatures.

4. The computer-implemented method of claim 1 , wherein the correlating the one or more medical descriptions extracted from the textual medical information related to the subject with the characteristics of the image of the medical condition of the subject to generate the subject signature comprises:

analyzing the textual medical information related to the subject to extract the one or more medical descriptions of the medical condition of the subject.

5. The computer-implemented method of claim 1 , wherein the comparing the subject signature with a plurality of reference signatures comprises:

calculating a match score between the subject signature and each of the plurality of reference signatures; and

determining a closest matching reference signature corresponding to the subject signature based on the match scores.

6. The computer-implemented method of claim 5 , wherein:

the characteristics of the image of the medical condition of the subject and the characteristics of the image of the corresponding medical condition include respective feature vectors; and

the calculating a match score between the subject signature and each of the plurality of reference signatures further comprises:

comparing the feature vector of the image of the medical condition of the subject with a respective second feature vector of each of the images of the corresponding medical conditions.

7. The computer-implemented method of claim 5 , wherein the match score for a reference signature is based on a distance between the reference signature and the subject signature.

8. A computer program product comprising:

a computer readable storage medium having computer readable program code embodied therewith for execution on a processing system, the computer readable program code being configured to be executed by the processing system to:

correlate one or more medical descriptions extracted from textual medical information of a subject with characteristics of an image of a medical condition of the subject to generate a subject signature, the correlating further comprising the processing system being configured to:

analyze the image of the medical condition of the subject to produce a feature vector representing features found in the image,

produce a vector based on the one or more medical descriptions, each element of the vector having a first value representing a respective detectable image feature that is included in the one or more medical descriptions, and each element of the vector having a second value representing a respective detectable image feature that is not included in the one or more medical descriptions, and

generate the subject signature based on the vector and the feature vector;

compare the subject signature with a plurality of reference signatures to determine a closest matching reference signature to the subject signature, wherein each reference signature is based on a corresponding second vector produced based on one or more reference medical descriptions extracted from textual medical information associated with a corresponding medical condition and a second feature vector representing features found in an image of the corresponding medical condition;

identify the medical condition of the subject based on the corresponding medical condition associated with the determined closest matching reference signature; and

provide information regarding the identified medical condition of the subject.

9. The computer program product of claim 8 , wherein the medical condition includes a disease.

10. The computer program product of claim 8 , wherein each of the reference signatures is generated by:

performing analytics on each of a plurality of textual medical information corresponding to a plurality of medical conditions to extract respective one or more reference medical descriptions of each of the plurality of medical conditions;

analyzing images of each of the plurality of medical conditions to extract respective image characteristics of each of the plurality of medical conditions; and

correlating the extracted respective one or more reference medical descriptions of each of the plurality of medical conditions with the corresponding respective image characteristics to generate each of the reference signatures.

11. The computer program product of claim 8 , wherein the correlating the one or more medical descriptions extracted from the textual medical information of the subject with the characteristics of the image of the medical condition of the subject to generate the subject signature further comprises:

analyzing the textual medical information of the subject to extract the one or more medical descriptions of the medical condition of the subject.

12. The computer program product of claim 8 , wherein the comparing the subject signature with a plurality of reference signatures comprises;

calculating a match score between the subject signature and each of the plurality of reference signatures; and

determining the closest matching reference signature corresponding to the subject signature based on the match scores.

13. The computer program product of claim 12 , wherein:

the characteristics of the image of the medical condition of the subject and the characteristics of the image of the corresponding medical condition include respective feature vectors; and

the calculating a match score between the subject signature and each of the plurality of reference signatures further comprises:

comparing the feature vector of the image of the medical condition of the subject with a respective second feature vector of each of the images of the corresponding medical conditions.

14. The computer program product of claim 12 , wherein the match score for a reference signature is based on a distance between the reference signature and the subject signature.

15. A processing device comprising:

at least one processor;

a memory; and

a communication bus connecting the at least one processor with the memory, wherein the memory has stored therein instructions, which when executed by the at least one processor cause the processing device to perform a method comprising:

correlating one or more medical descriptions extracted from textual medical information related to a subject with characteristics of an image of a medical condition of the subject to generate a subject signature, the correlating further comprising:

analyzing the image of the medical condition of the subject to produce a feature vector representing features found in the image,

producing a vector based on the one or more medical descriptions, each element of the vector having a first value representing a respective detectable image feature that is included in the one or more medical descriptions, and each element of the vector having a second value representing a respective detectable image feature that is not included in the one or more medical descriptions, and

generating the subject signature based on the vector and the feature vector;

comparing the subject signature with a plurality of reference signatures to determine a closest matching reference signature to the subject signature, wherein each reference signature is based on a corresponding second vector produced based on one or more reference medical descriptions extracted from textual medical information associated with a corresponding medical condition and a second feature vector representing features found in an image of the corresponding medical condition;

identifying the medical condition of the subject based on the corresponding medical condition associated with the determined closest matching reference signature; and

providing information regarding the identified medical condition of the subject.

16. The processing device of claim 15 , wherein the medical condition includes a disease.

17. The processing device of claim 15 , further comprising:

generating the plurality of reference signatures, the generating the plurality of reference signatures further comprising:

performing analytics on a plurality of textual medical information corresponding to a plurality of medical conditions to extract respective one or more reference medical descriptions of each of the plurality of medical conditions;

analyzing images of the plurality of medical conditions to extract respective image characteristics of each of the plurality of medical conditions; and

correlating the extracted respective one or more reference medical descriptions of the plurality of medical conditions with the corresponding respective image characteristics to generate the reference signatures.

18. The processing device of claim 15 , wherein the correlating to generate the subject signature comprises:

analyzing the textual medical information related to the subject to extract the one or more medical descriptions of the medical condition of the subject.

19. The processing device of claim 15 , wherein the comparing the subject signature with a plurality of reference signatures comprises:

calculating a match score between the subject signature and each of the plurality of reference signatures; and

determining the closest matching reference signature corresponding to the subject signature based on the match scores.

20. The processing device of claim 19 , wherein:

the characteristics of the image of the medical condition of the subject and the characteristics of the image of the corresponding medical condition include respective feature vectors; and

the calculating a match score between the subject signature and each of the plurality of reference signatures further comprises:

comparing the feature vector of the image of the medical condition of the subject with a respective second feature vector of each of the images of the corresponding medical conditions.