# US 12,217,425 B2 · Method for discriminating suspicious lesion in medical image, method for interpreting medical image, and computing device implementing the methods > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 12,217,425 B2 - Document ID: US12217425B2 - Status: Granted - Kind: B2 - Application number: 18/584,049 - Filed: 2024-02-22 - Published: 2024-06-13 - Granted: 2025-02-04 - Earliest claimed: 2019-05-24 - Assignees: Lunit Inc. - Applicants: Lunit Inc. - Inventors: Hyo-Eun Kim; Hyeonseob Nam - Web page: https://patentgrep.com/patent/US12217425B2 ## CPC - G06T 7/0012: {Biomedical image inspection} - G06F 18/2431: Multiple classes - G06F 18/24323: {Tree-organised classifiers} - G06N 3/0464: Convolutional networks [CNN, ConvNet] - G06N 3/08: Learning methods - G06N 3/09: Supervised learning - G06N 20/00: Machine learning - G06V 10/40: Extraction of image or video features - G06V 10/764: using classification, e.g. of video objects - G06N 3/045: Combinations of networks - G06T 2207/20081: Training; Learning - G06T 2207/20084: Artificial neural networks [ANN] - G06T 2207/30068: Mammography; Breast - G06T 2207/30096: Tumor; Lesion - G06V 2201/03: Recognition of patterns in medical or anatomical images - G06T: IMAGE DATA PROCESSING OR GENERATION, IN GENERAL - G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N) - G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS - G06V: IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING ## Abstract A method for interpreting an input image by a computing device operated by at least one processor is provided. The method for interpreting an input image comprises storing an artificial intelligent (AI) model that is trained to classify a lesion detected in the input image as suspicious or non-suspicious and, under a condition of being suspicious, to classify the lesion detected in the input image as malignant or benign-hard representing that the lesion is suspicious but determined to be benign, receiving an analysis target image, by using the AI model, obtaining a classification class of a target lesion detected in the analysis target image and, when the classification class is the suspicious, obtaining at least one of a probability of being suspicious, a probability of being benign-hard, and a probability of malignant for the target lesion, and outputting an interpretation result including at least one probability obtained for the target lesion. ## Claims ### Claim 1 A computing device comprising: a non-transitory memory storage that stores instructions; and a processor that executes the instructions, wherein the processor is configured to: receive a target image, using an artificial intelligent (AI) model that is trained to classify a lesion detected in a medical image into one of predefined types of lesions, obtain a classification class of a target lesion detected in the target image among the predefined types of lesions, and output an interpretation result including the classification class and classification probability of the target lesion, the classification probability indicating a probability that the target lesion is classified into the classification class by the AI model. ### Claim 2 (depends on 1) The computing device of claim 1, wherein training data for training the AI model consists of each medical image and a label annotated with classification class for at least one lesion included in corresponding medical image. ### Claim 3 (depends on 1) The computing device of claim 1, wherein the classification class includes types of lesions with different degrees of classification learning difficulties of the AI model. ### Claim 4 (depends on 1) The computing device of claim 1, wherein the classification class includes types of lesions that have different degrees of difficulty for reading to a doctor. ### Claim 5 (depends on 1) The computing device of claim 1, wherein the classification class includes at least one of a non-suspicious class, a suspicious class, a normal class, a benign class, a malignant class, a benign-hard class, or a benign-easy class. ### Claim 6 (depends on 1) The computing device of claim 1, wherein the interpretation result is displayed according to the classification class. ### Claim 7 (depends on 1) The computing device of claim 1, wherein the interpretation result includes a location of the target lesion and a probability of being malignant, when the target lesion is classified as a malignant class. ### Claim 8 (depends on 1) The computing device of claim 1, wherein the interpretation result includes information indicating that the target lesion is a benign-hard lesion, and a probability of being suspicious or a probability of being benign-hard, when the target lesion is classified as a suspicious class and a benign class. ### Claim 9 (depends on 1) The computing device of claim 1, wherein the interpretation result includes information indicating that the target lesion is a non-suspicious lesion and a probability of being non-suspicious, when the target lesion is classified as a non-suspicious class. ### Claim 10 (depends on 1) The computing device of claim 1, wherein the processor is configured to output an interpretation result including information that no lesion is detected in the target image, when the target image is classified as a normal class. ### Claim 11 (depends on 1) The computing device of claim 1, wherein the interpretation result further includes a location of the target lesion detected in the target image. ### Claim 12 A method for interpreting an input image by a computing device, the method comprising: receiving a target image; using an artificial intelligent (AI) model that is trained to classify a lesion detected in a medical image into one of predefined types of lesions, obtaining a classification class of a target lesion detected in the target image among the predefined types of lesions; and outputting an interpretation result including the classification class and classification probability of the target lesion, the classification probability indicating a probability that the target lesion is classified into the classification class by the AI model. ### Claim 13 A non-transitory computer-readable storage medium, storing a program comprising instructions for causing a computing device to execute to: receive a target image; using an artificial intelligent (AI) model that is trained to classify a lesion detected in a medical image into one of predefined types of lesions, obtain a classification class of a target lesion detected in the target image among the predefined types of lesions; and output an interpretation result including the classification class and classification probability of the target lesion, the classification probability indicating a probability that the target lesion is classified into the classification class by the AI model.