# US 11,998,318 B2 · System and method of using visually-descriptive words to diagnose ear pathology > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 11,998,318 B2 - Document ID: US11998318B2 - Status: Granted - Kind: B2 - Application number: 17/545,681 - Filed: 2021-12-08 - Published: 2022-08-18 - Granted: 2024-06-04 - Earliest claimed: 2021-02-12 - Assignees: Ohio State Innovation Foundation - Applicants: Ohio State Innovation Foundation - Inventors: Metin Gurcan; Aaron Moberly - Web page: https://patentgrep.com/patent/US11998318B2 ## CPC - A61B 5/126: {measuring compliance or mechanical impedance of the tympanic membrane} - A61B 5/7264: {Classification of physiological signals or data, e.g. using neural networks, statistical classifiers, expert systems or fuzzy systems} - G06N 3/08: Learning methods - G16H 15/00: ICT specially adapted for medical reports, e.g. generation or transmission thereof - G16H 50/50: for simulation or modelling of medical disorders - G16H 50/70: for mining of medical data, e.g. analysing previous cases of other patients - G06F 40/20: Natural language analysis (semantic analysis of natural language G06F40/30) - G06F 40/30: Semantic analysis - G10L 15/08: Speech classification or search - G16H 50/20: for computer-aided diagnosis, e.g. based on medical expert systems - A61B: DIAGNOSIS; SURGERY; IDENTIFICATION (usefulness limited to only animals A61D) - G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS - G16H: HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA - G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N) - G10L: SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING ## Abstract Disclosed herein are systems and methods to detect a wide range of eardrum conditions by using visually-descriptive words of a tympanic membrane of a subject. ## Claims ### Claim 1 A method of classifying tympanic membrane pathologies from descriptive words, comprising: receiving one or more descriptive words describing a condition of a tympanic membrane of a subject; and determining, by a computing device, using a model comprised of a plurality of tympanic membrane conditions and corresponding word clouds, wherein each word cloud is comprised of words most frequently associated with each one of the plurality of tympanic membrane conditions as determined by an artificial intelligence algorithm, a diagnosis of the tympanic membrane of the subject based on the received one or more descriptive words describing the condition of the tympanic membrane of the subject, wherein the model is developed by obtaining a plurality of images of tympanic membranes having the plurality of tympanic membrane conditions, forming a composite image of a tympanic membrane for each of the plurality of tympanic membrane conditions, having a specialist provide a descriptive report of each composite image and its corresponding tympanic membrane condition; and analyzing using the artificial intelligence algorithm each specialist descriptive report to determine the word cloud of words most frequently associated with the corresponding tympanic membrane image, and wherein one or more natural language processing (NLP) tools were used to preprocess each specialist descriptive report, said NLP tools comprising one or more of tokenization, stop word removal, word normalization, and punctuation erasing. ### Claim 2 (depends on 1) The method of claim 1, wherein the one or more descriptive words are received electronically as a voice of a healthcare provider, or wherein the one or more descriptive words are received electronically as text, and wherein the one or more descriptive words are received in real-time, or wherein the one or more descriptive words are recorded and received later in time. ### Claim 3 (depends on 1) The method of claim 1, wherein a bag-of-words (BoW) model (also known as a term-frequency counter) was applied to each preprocessed specialist descriptive report to obtain a number of times that particular words appear in each report, implying their frequency. ### Claim 4 (depends on 1) The method of claim 1, wherein once the model is trained, the model can automatically create a word cloud characterizing additional tympanic membrane images. ### Claim 5 (depends on 1) The method of claim 1, wherein determining, the diagnosis of the tympanic membrane of the subject based on the received one or more descriptive words describing the condition of the tympanic membrane of the subject comprises comparing the received one or more descriptive words describing the condition of the tympanic membrane to the word clouds using neighborhood components analysis (NCA), where NCA learns a linear transformation of data that maximizes k-nearest neighbor performance. ### Claim 6 (depends on 1) The method of claim 1, wherein the model comprises a database that associates each of the plurality of tympanic membrane conditions with its corresponding word cloud. ### Claim 7 (depends on 1) The method of claim 1, wherein the diagnosis of the tympanic membrane of the subject is one of normal, effusion, retraction, or tympanosclerosis. ### Claim 8 A system for classifying tympanic membrane pathologies from descriptive words, comprising: a device configured to view a tympanic membrane of a subject; and a processor in communication with a memory, wherein the memory contains computer-executable instructions that when executed by the processor cause the processor to: receive one or more descriptive words describing a condition of the tympanic membrane of a subject; and determine, using a model comprised of a plurality of tympanic membrane conditions and corresponding word clouds, wherein each word cloud is comprised of words most frequently associated with each one of the plurality of tympanic membrane conditions as determined by an artificial intelligence algorithm, a diagnosis of the tympanic membrane of the subject based on the received one or more descriptive words describing the condition of the tympanic membrane of the subject, wherein the model is developed by obtaining a plurality of images of tympanic membranes having the plurality of tympanic membrane conditions, forming a composite image of a tympanic membrane for each of the plurality of tympanic membrane conditions, having a specialist provide a descriptive report of each composite image and its corresponding tympanic membrane condition; and analyzing using the artificial intelligence algorithm each specialist descriptive report to determine the word cloud of words most frequently associated with the corresponding tympanic membrane image, and wherein one or more natural language processing (NLP) tools were used to preprocess each specialist descriptive report, said NLP tools comprising one or more of tokenization, stop word removal, word normalization, and punctuation erasing. ### Claim 9 (depends on 8) The system of claim 8, wherein the one or more descriptive words are received electronically as a voice of a healthcare provider or wherein the one or more descriptive words are received electronically as text, and wherein the one or more descriptive words are received in real-time, or wherein the one or more descriptive words are recorded and received later in time. ### Claim 10 (depends on 8) The system of claim 8, wherein a bag-of-words (BoW) model (also known as a term-frequency counter) was applied to each preprocessed specialist descriptive report to obtain a number of times that particular words appear in each report, implying their frequency. ### Claim 11 (depends on 8) The system of claim 8, wherein once the model is trained, the model can automatically create a word cloud characterizing additional tympanic membrane images. ### Claim 12 (depends on 8) The system of claim 8, wherein the processor executing computer-executable instructions that cause the processor to determine the diagnosis of the tympanic membrane of the subject based on the received one or more descriptive words describing the condition of the tympanic membrane of the subject comprises the processor executing computer-readable instructions that cause the processor to compare the received one or more descriptive words describing the condition of the tympanic membrane to the word clouds using neighborhood components analysis (NCA), where NCA learns a linear transformation of data that maximizes k-nearest neighbor performance. ### Claim 13 (depends on 8) The system of claim 8, wherein the model comprises a database stored in the memory that associates each of the plurality of tympanic membrane conditions with its corresponding word cloud. ### Claim 14 (depends on 8) The system of claim 8, wherein the diagnosis of the tympanic membrane of the subject is one of normal, effusion, retraction, or tympanosclerosis. ### Claim 15 A non-transitory computer-program product comprising computer executable code sections stored on a computer-readable medium, said computer executable code sections for performing a method of classifying tympanic membrane pathologies from descriptive words, comprising: receiving one or more descriptive words describing a condition of a tympanic membrane of a subject; and determining, using a model comprised of a plurality of tympanic membrane conditions and corresponding word clouds, wherein each word cloud is comprised of words most frequently associated with each one of the plurality of tympanic membrane conditions as determined by an artificial intelligence algorithm, a diagnosis of the tympanic membrane of the subject based on the received one or more descriptive words describing the condition of the tympanic membrane of the subject, wherein the model is developed by obtaining a plurality of images of tympanic membranes having the plurality of tympanic membrane conditions, forming a composite image of a tympanic membrane for each of the plurality of tympanic membrane conditions, having a specialist provide a descriptive report of each composite image and its corresponding tympanic membrane condition; and analyzing using the artificial intelligence algorithm each specialist descriptive report to determine the word cloud of words most frequently associated with the corresponding tympanic membrane image, and wherein one or more natural language processing (NLP) tools were used to preprocess each specialist descriptive report, said NLP tools comprising one or more of tokenization, stop word removal, word normalization, and punctuation erasing. ### Claim 16 (depends on 15) The computer-program product of claim 15, wherein the one or more descriptive words are received electronically as a voice of a healthcare provider or wherein the one or more descriptive words are received electronically as text, and wherein the one or more descriptive words are received in real-time, or wherein the one or more descriptive words are recorded and received later in time. ### Claim 17 (depends on 15) The computer-program product of claim 15, wherein a bag-of-words (BoW) model (also known as a term-frequency counter) was applied to each preprocessed specialist descriptive report to obtain a number of times that particular words appear in each report, implying their frequency. ### Claim 18 (depends on 15) The computer-program product of claim 15, wherein once the model is trained, the model can automatically create a word cloud characterizing additional tympanic membrane images. ### Claim 19 (depends on 15) The computer-program product of claim 15, wherein determining, the diagnosis of the tympanic membrane of the subject based on the received one or more descriptive words describing the condition of the tympanic membrane of the subject comprises comparing the received one or more descriptive words describing the condition of the tympanic membrane to the word clouds using neighborhood components analysis (NCA), where NCA learns a linear transformation of data that maximizes k-nearest neighbor performance. ### Claim 20 (depends on 15) The computer-program product of claim 15, wherein the model comprises a database that associates each of the plurality of tympanic membrane conditions with its corresponding word cloud. ### Claim 21 (depends on 15) The computer-program product of claim 15, wherein the diagnosis of the tympanic membrane of the subject is one of normal, effusion, retraction, or tympanosclerosis.