Endoscopy video feature enhancement platform
An endoscopy video feature enhancement platform (EVFEP) is connected to the output of any type endoscope system, and inputs and captures the output video. The video is visually augmented live with indicators of possible polyp detection and localization, polyp attributes, and procedure metrics, based on the collective learning of the output results of many different types of endoscopy systems on a large scale. An artificial intelligence model is trained on confirmed polyp detection previously determined by this and other EVFEP devices used with many different types of endoscope systems on a large scale. Augmented video, images and automatically generated short video clips of key procedure segments are passed to a reporting system, and supplemented with meta data.
1. An endoscopy video feature enhancement platform, comprising:
an input video interface to input an unaugmented mirrored monitor video stream output from an endoscopy system directly through a mirrored part on the endoscopy system;
a digital video splitter to digitally split the input unaugmented mirrored monitor video stream into both an unaugmented mirrored monitor video stream for output to an unaugmented video interface, and a mirrored endoscopy video stream for post-endoscopy-system real-time augmentation;
a detection module to apply in real time a previously trained detection model to the mirrored endoscopy video stream, and generate in real time detection of a region of at least one possible abnormality in a subject patient shown in a relevant portion of the mirrored endoscopy video stream;
an abnormality attribute collection module to determine in real time at least one attribute of the at least one possible abnormality generated by the detection module, to overlay in real time a visual border around the detected region within each frame including the detected region, and to overlay in real time the at least one attribute of the at least one possible abnormality within each said frame, of the relevant portion of the mirrored endoscopy video stream, to generate an augmented endoscopy video stream;
a video augmentation module to overlay in real time, within each said frame of the augmented endoscopy video stream including the visual border around the detected region, a visual indicator of a location of the at least one possible abnormality within each said frame, and the at least one attribute of the at least one possible abnormality;
an output video interface; and
a video output device to output the augmented endoscopy video stream to the output video interface.
2. The endoscopy video feature enhancement platform according to claim 1 , wherein:
the input video interface is a Universal Serial Bus interface.
3. The endoscopy video feature enhancement platform according to claim 1 , wherein:
the output video interface is an HDMI output.
4. The endoscopy video feature enhancement platform according to claim 1 , further comprising:
a video converter to convert a first video format of the input unaugmented mirrored monitor video stream to a different, second video format.
5. The endoscopy video feature enhancement platform according to claim 1 , wherein:
the video augmentation module further overlays on the augmented endoscopy video stream textual information relating to the at least one attribute of the at least one possible abnormality.
6. A method of augmenting an endoscopy video feed output from an endoscope system, comprising:
inputting an unaugmented mirrored monitor video stream output from an endoscopy system directly through a mirrored part on the endoscopy system or optionally through a digital video splitter;
digitally splitting the input unaugmented mirrored monitor video stream into both an unaugmented mirrored monitor video stream for output to an unaugmented video interface, and a mirrored endoscopy video stream for post-endoscopy-system real-time augmentation;
applying in real time a previously trained detection model to the mirrored endoscopy video stream;
generating in real time detection of a region of at least one possible abnormality in a subject patient shown in a relevant portion of the mirrored endoscopy video stream;
determining in real time at least one attribute of the at least one possible abnormality in the detected region;
overlaying in real time a visual border around the detected region within each frame including the detected region, and to overlay in real time the at least one attribute of the at least one possible abnormality within each said frame, of the relevant portion of the mirrored endoscopy video stream, to generate an augmented endoscopy video stream;
overlaying in real time, within each said frame of the augmented endoscopy video stream including the visual border around the detected region, a visual indicator of a location of the at least one possible abnormality within each said frame, and the at least one attribute of the at least one possible abnormality; and
outputting the augmented endoscopy video stream to an output video interface.
7. The method of augmenting an endoscopy video feed output from an endoscope system according to claim 6 , further comprising:
converting a first video format of the input unaugmented mirrored monitor video stream to a different, second video format.
8. The method of augmenting an endoscopy video feed output from an endoscope system according to claim 6 , further comprising:
overlaying in real time on the augmented endoscopy video stream textual information relating to the at least one attribute of the at least one possible abnormality.
9. The method of augmenting an endoscopy video feed output from an endoscope system according to claim 6 , wherein:
the input unaugmented mirrored monitor video stream output by the endoscopy system is already enhanced by the endoscope system.
10. The method of augmenting an endoscopy video feed output from an endoscope system according to claim 6 , further comprising:
periodically updating the trained detection model with a new detection model trained on additional endoscopy results.
11. An endoscopy video feature enhancement platform, comprising:
an input video interface to input an unaugmented mirrored monitor video stream output from an endoscopy system directly through a digital video splitter to digitally split the input unaugmented mirrored monitor video stream into both an unaugmented mirrored monitor video stream for output to an unaugmented video interface, and a mirrored endoscopy video stream for post-endoscopy-system real-time augmentation;
a detection module to apply in real time a previously trained detection model to the mirrored endoscopy video stream, and generate in real time detection of a region of at least one possible abnormality in a subject patient shown in a relevant portion of the mirrored endoscopy video stream;
an abnormality attribute collection module to determine in real time at least one attribute of the at least one possible abnormality generated by the detection module, to overlay in real time a visual border around the detected region within each frame including the detected region, and to overlay in real time the at least one attribute of the at least one possible abnormality within each said frame, of the relevant portion of the mirrored endoscopy video stream, to generate an augmented endoscopy video stream;
a video augmentation module to overlay in real time, within each said frame of the augmented endoscopy video stream including the visual border around the detected region, a visual indicator of a location of the at least one possible abnormality within each said frame, and the at least one attribute of the at least one possible abnormality;
an output video interface; and
a video output device to output the augmented endoscopy video stream to the output video interface.