Endoscopic device, control method and computer program for the endoscopic device
AssigneeMedintech Inc.
InventorsChi Won Lee, Myung Joon Kim, Tae Hyeon Kim, Seok Ho Cho, Dong Hyeok Kim
Life of this patent
Estimating the term…
Abstract
A control method for an endoscopic device includes obtaining, from an image sensor, an image of an inside of a body, inputting the image to an artificial neural network model trained to classify the obtained image, classifying and outputting, by the artificial neural network model, the image as an air situation image, a water situation image, and/or a suction situation image, and controlling the endoscopic device to drive an air unit, a water unit, and/or a suction unit according to an output classification result.
Claims12
A control method for an endoscopic device comprising at least one processor, the control method comprising:
- obtaining, from an image sensor, an image of an inside of a body;
- inputting the image to an artificial neural network model configured to classify, for each frame, the image of the inside of the body as at least one of an air situation image, a water situation image, or a suction situation image;
- performing post-classification processing comprising:
- (i) applying a classification-specific threshold to the classified image, and
- (ii) determining a matching degree with classification results of a previous frame and a subsequent frame of the classified image;
- and
- controlling the endoscopic device to automatically open or close a corresponding valve of at least one of an air unit, a water unit, or a suction unit of the endoscopic device in response to the classification satisfying the classification-specific threshold and the matching degree exceeding a criterion, thereby executing at least one of airing, watering, or suction through a working channel of the endoscopic device,
- wherein the post-classification processing is configured to enhance a classification accuracy of at least one of the air situation image, the water situation image, or the suction situation image, and
- wherein the classification-specific threshold is variously set to a first threshold value, a second threshold value, or a third threshold value, so that airing, watering, and sucking up are differently controlled according to situations by applying different threshold values depending on a condition of a patient during an endoscopic procedure.
An endoscopic device comprising:
- a memory storing an image of an inside of a body captured by the endoscopic device;
- and
- a processor configured to:
- input the image of the inside of the body to a trained artificial neural network model,
- classify, for each frame, the image of the inside of the body as at least one of an air situation image, a water situation image, or a suction situation image,
- perform post-classification processing comprising:
- (i) applying a classification-specific threshold to the classified image, and
- (ii) determining a matching degree with classification results of a previous frame and a subsequent frame of the classified image;
- and
- control, via a driver, at least one of an air unit, a water unit, or a suction unit by automatically opening or closing a corresponding valve in response to the classification satisfying the classification-specific threshold and the matching degree exceeding a criterion, so as to execute at least one of airing, watering, or suction through a working channel of the endoscopic device,
- wherein the post-classification processing is configured to enhance a classification accuracy of at least one of the air situation image, the water situation image, or the suction situation image, and
- wherein the classification-specific threshold is configured to be variously set to a first threshold value, a second threshold value, or a third threshold value, so that airing, watering, and sucking up are differently controlled according to situations by applying different threshold values depending on a condition of a patient during an endoscopic procedure.
Dependencies are inferred from the claim text.