De-identifying data obtained from microphones
An aspect includes a computer-implemented method that de-identifies data received from microphones. The method includes receiving data from one or more microphones and de-identifying the data. The de-identifying includes inputting the data into a machine learning system that has been trained to detect patterns in the data that are likely to identify a specific entity, and to remove the detected patterns from the data to generate de-identified data. An output from the machine learning system is received, where the output includes the de-identified data. According to some aspects, the microphones can be located in an operating room.
1. A computer-implemented method comprising:
receiving data, at a computing system, from one or more microphones in an operating room;
de-identifying the data by the computing system, the de-identifying comprising:
inputting the data into a machine learning system that has been trained to detect patterns in the data that are likely to identify a specific entity, and executing a de-identification inference engine to remove the detected patterns from the data to generate de-identified data comprising a transformation of one or more phrases in the data based on a combination of keywords and actions defined for a surgical environment;
receiving an output from the machine learning system, the output including the de-identified data; and
providing the output to a data collection system.
2. The method of claim 1 , further comprising one or both of training and retraining the machine learning system.
3. The method of claim 1 , wherein the received data is in an audio format and the method further comprises:
converting the received data into a discrete format; and
inputting the data in the discrete format into the machine learning system.
4. The method of claim 1 , wherein the de-identified data is in a text format.
5. The method of claim 4 , further comprising converting the de-identified data into an audio format.
6. The method of claim 1 , wherein the specific entity is a person.
7. The method of claim 6 , wherein at least one of the patterns includes a phrase that is associated with the person.
8. The method of claim 1 , wherein at least one of the patterns identifies a group that includes two or more entities.
9. The method of claim 1 , further comprising providing the de-identified data to a second machine learning system comprising one or more machine learning models that are trained to identify a plurality of surgical phases in a video of a surgical procedure.
10. A computer program product comprising a memory device having computer executable instructions stored in a non-transitory form thereon, which when executed by one or more processors cause the one or more processors to perform operations comprising:
receiving labeled training data comprising a plurality of sets of phrases, each set of phrases comprising a phrase and a corresponding de-identified phrase; and
generating, using the labeled training data, a machine learning model to detect patterns in input data that are likely to identify a specific entity and to remove the detected patterns from the data as a transformation of one or more phrases in the data based on a combination of keywords and actions defined for a surgical environment.
11. The computer program product of claim 10 , wherein the generating the machine learning model is further based at least in part on a list comprising identifying phrases that each identify one or more entities.
12. The computer program product of claim 10 , wherein the operations further comprise retraining the machine learning model.
13. The computer program product of claim 10 , wherein the operations further comprise validating the machine learning model.
14. The computer program product of claim 10 , wherein the input data is received from one or more microphones located in an operating room.
15. The computer program product of claim 10 , wherein the operations further comprise:
receiving data from one or more microphones;
de-identifying the data, the de-identifying comprising inputting the data into the machine learning model; and
receiving an output from the machine learning system, the output including the de-identified data.
16. A system comprising:
a data collection system comprising one or more storage devices configured to capture a video of a surgical procedure, the video including data received from one or more microphones in an operating room;
a machine learning execution system configured to execute one or more machine-learning models on a processing system to identify a plurality of surgical phases in the video and to de-identify the data received from the one or more microphones by performing a transformation of one or more phrases in the data based on a combination of keywords and actions defined for a surgical environment; and
an output generator configured to store the video with the surgical phases identified and the de-identified data.
17. The system of claim 16 , wherein the data from the one or more microphones is converted into a text format prior to being input to the model learning execution system and the de-identified data is in the text format.
18. The system of claim 16 , wherein the de-identifying of the data comprises removing any patterns in the data that are likely to identify a specific entity.
19. The system of claim 16 , wherein the de-identified data is in an audio format.