Verb Surgical Inc.

Meysam Torabi

Richard William Timm

Jocelyn Elaine Barker

Varun Kejriwal Goel

Pablo E Garcia Kilroy

Danyal Fer

Amer Ghanem

Rami Abukhalil

User input device (UID) disengagement-detection techniques based on real-time time-series data processing and deep learning. A long short-term memory (LSTM) network-based classifier is trained based on the acquired time-series data of UID motions including both surgical motions and docking motions. The classifier is then used during teleoperation sessions to monitor the movements of UIDs, and continuously classify the real-time UID motions as either teleoperation motions or docking motions. The disengagement-detection techniques can immediately disengage the UIDs from the surgical tools as soon as the monitored UID motions are classified as docking motions, thereby preventing unintended surgical tool motions. The disclosed disengagement-detection techniques allow the UIDs and the surgical tools to become disengaged naturally by simply having the user place the UIDs back to their docking positions without having to take any additional actions. Other aspects are also described and claimed.

Deep disengagement detection during telesurgery

Embodiments described herein include a process for detecting energy tool activations. The process can begin by receiving a surgical video of a surgical procedure involving energy tool activations. The process then applies a sequence of sampling windows to the surgical video to generate a sequence of windowed samples of the surgical video. Next, for each windowed sample in the sequence of windowed samples, the process applies a deep-learning model to a sequence of video frames within the windowed sample to generate an activation/non-activation inference and a confidence level associated with the activation/non-activation inference for the windowed sample. As a result, a sequence of activation/non-activation inferences and a sequence of associated confidence levels are generated. The process subsequently identifies a sequence of activation events in the surgical video based on the sequence of activation/non-activation inferences and the sequence of associated confidence levels.

Energy tool activation detection in surgical videos using deep learning

Disclosed are various face-detection and human de-identification systems and techniques based on deep learning. In one aspect, a process for de-identifying people captured in an operating room (OR) video is disclosed. This process can begin by receiving a sequence of video frames from an OR video. Next, the process applies a first machine-learning face detector based on a first deep-learning model to each video frame in the sequence of video frames to generate a first set of detected faces. The process further applies a second machine-learning face detector to the sequence of video frames to generate a second set of detected faces, wherein the second machine-learning face detector is constructed based on a second deep-learning model different from the first deep-learning model. The process subsequently de-identifies the received sequence of video frames by blurring out both the first set of detected faces and the second set of detected faces.

Robust operating room video anonymization based on ensemble deep learning

Embodiments described herein provide various techniques and systems for building machine-learning surgical tool presence/absence detection models for processing surgical videos and predicting whether a surgical tool is present or absent in each video frame of a surgical video. In one aspect, a process for ensuring patient safety during a laparoscopic or robotic surgery involving an energy tool is disclosed. The process can begin receiving a real-time control signal indicating an operating state of an energy tool during the surgery. Next, the process receives real-time endoscope video images of the surgery. The process simultaneously applies a machine-learning surgical tool presence/absence detection model to the real-time endoscope video images to generate real-time decisions on a location of the energy tool in the real-time endoscope video images. The process then checks the real-time control signal against the real-time decisions to identify an unsafe event and takes a proper action when an unsafe event is identified.

Real-time surgical tool presence/absence detection in surgical videos

A method for disengagement detection of a surgical instrument of a surgical robotic system, the method comprising: determining whether a user's head is unstable prior to disengagement of a teleoperation mode; determining whether a pressure release has occurred relative to at least one of a first user input device or a second user input device for controlling a surgical instrument of the surgical robotic system during the teleoperation mode; and in response to determining the user's head is unstable or determining the pressure release has occurred, determining whether a distance change between the first user input device and the second user input device indicates the user is performing an unintended action prior to disengagement of the teleoperation mode.

Surgeon disengagement detection during termination of teleoperation

Various user-presence/absence detection techniques based on deep learning are provided. These user-presence/absence detection techniques can include building/training a deep-learning model including a user-presence/absence classifier based on training images of a user-seating area of a surgeon console under various clinically-relevant conditions. The trained user-presence/absence classifier can then be used during teleoperation/surgical procedures to monitor/track users in the user-seating area of the surgeon console, and continuously classify captured real-time video images of the user-seating area into either a user-presence classification or a user-absence classification. In some embodiments, the user-presence/absence classifier can be used to detect a user-switching event at the surgeon console when a second user is detected to have entered the user-seating area after a first user is detected to have exited the user-seating area. If the second user is identified as a new user, this can trigger a recalibration procedure to recalibrate surgeon-console settings for the new user.

User switching detection during robotic surgeries using deep learning

Various user input device (UID) disengagement-detection techniques based on real-time time-series data processing and deep learning. More specifically, various UID disengagement-detection techniques include training a long short-term memory (LSTM) network-based classifier based on the acquired time-series data of UID motions including both surgical motions and docking motions. The trained deep-learning classifier can then be used during teleoperation sessions to monitor the movements of UIDs, and continuously classify the real-time UID motions as either teleoperation motions or docking motions. The disengagement-detection techniques can immediately disengage the UIDs from the surgical tools as soon as the monitored UID motions are classified as docking motions by the trained classifier, thereby preventing unintended surgical tool motions. The disengagement-detection techniques allow the UIDs and the surgical tools to become disengaged naturally by simply having the user putting the UIDs back to their docking positions without having to take any additional actions.

Various user-presence/absence recognition techniques based on deep learning are provided. More specifically, these user-presence/absence recognition techniques include building/training a CNN-based image recognition model including a user-presence/absence classifier based on training images collected from the user-seating area of a surgeon console under various clinically-relevant conditions/cases. The trained user-presence/absence classifier can then be used during teleoperation/surgical procedures to monitor/track users in the user-seating area of the surgeon console, and continuously classify the real-time video images of the user-seating area as either a user-presence state or a user-absence state. In some embodiments, the user-presence/absence classifier can be used to detect a user-switching event at the surgeon console when a second user is detected to have entered the user-seating area after a first user is detected to have exited the user-seating area. If the second user is identified as a new user, the disclosed techniques can trigger a recalibration procedure for the new user.

Growing community of inventors

Union City, CA, United States of America

Meysam Torabi