Abstract: 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.

Abstract: Embodiments described in this disclosure include a process for collecting energy tool usage data from surgical videos and using such data for post surgery analysis. The process can begin by receiving a plurality of surgical videos of a surgical procedure involving an energy tool. For each surgical video in the plurality of surgical videos, the process detects a set of activation events in the surgical video, wherein each detected activation event includes an identified starting timestamp and a duration. The process further extracts a set of energy tool usage data based on the set of detected activation events, and then stores the extracted set of energy tool usage data in a database indexed based on a set of energy tool usage metrics. Next, in response to a user search request, the process returns the stored energy tool usage data that matches the search request from the database.