Phase-Based Surgical Video Segmentation for Machine-Learning Targets

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing surgical video analysis systems fail to effectively mine machine learning targets from a large cache of surgical videos to evaluate and improve surgical outcomes and skills, despite the availability of a vast amount of recorded surgical procedures.

Innovation Solution

A surgical video analysis system that segments surgical procedures into predefined phases, identifies clinical needs, and translates these needs into machine learning targets, establishing associative relationships among them to create searchable databases for evaluating surgeon skills and surgery quality.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If surgical videos are processed and analyzed manually, then analysis accuracy can be maintained, but processing time and labor costs increase significantly

Engineering Contradiction:
Improveanalysis accuracyVSAvoidprocessing time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The surgical video is segmented into multiple phases based on predefined phase definitions. Each phase represents a distinct surgical step or stage, allowing the system to process and analyze specific segments independently rather than analyzing the entire video as one continuous stream, thus reducing overall processing time while maintaining accuracy within each phase

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

Phase definitions are pre-established before video analysis begins. These predefined phases include phase identifiers, start/end time indicators, and associated machine learning targets. By having this structural framework prepared in advance, the system can quickly map video content to appropriate phases without performing complex real-time decision-making, thereby reducing processing time

Inventive Principle:
Principle #10Preliminary action

2Ease of manufacture

If machine learning models are trained on raw surgical videos without phase segmentation, then model development is simpler, but the precision of skill evaluation and outcome assessment decreases

Engineering Contradiction:
Improvemodel development simplicityVSAvoidskill evaluation precision
Core Design Contradiction:
Ease of manufactureVSMeasurement precision

Solution Approach 1:

The video is divided into phased segments, each associated with specific machine learning targets. This segmentation allows the system to evaluate surgeon skills and surgical outcomes with greater precision by analyzing specific phases rather than treating all video content uniformly, while still using standardized phase definitions that simplify the overall modeling process

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The system transforms raw video data into structured phase-based representations with associated machine learning targets. By changing the parameter representation from unstructured video streams to structured phase segments with defined characteristics, the system achieves more precise skill evaluation while maintaining manageable model complexity

Inventive Principle:
Principle #35Parameter changes

3Ease of operation

If all surgical video data is stored in a single database, then data retrieval is simpler, but search efficiency and data organization decrease

Engineering Contradiction:
Improvedata retrieval simplicityVSAvoidsearch efficiency
Core Design Contradiction:
Ease of operationVSProductivity

Solution Approach 1:

The database is segmented into multiple specialized databases, each storing specific types of information: phase definitions, video segments, machine learning targets, and results. This segmentation improves search efficiency by allowing queries to target specific database types rather than searching through all data, while maintaining organized access through the phase-based structure

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

Phase definitions serve as an intermediary layer between raw video data and analysis queries. This intermediary structure enables efficient searching by providing a standardized framework that links video content to specific phases and machine learning targets, improving both retrieval simplicity and search efficiency

Inventive Principle:
Principle #24Intermediary (Mediator)

Data Source

PatentEP3796827B1Machine-learning-oriented surgical video analysis system
Publication Date: 2025.10.29 VERB SURGICAL INC
  • EP3796827B1 patent drawingFigure 1
  • EP3796827B1 patent drawingFigure 2
  • EP3796827B1 patent drawingFigure 3

AI summary

Embodiments described herein provide various examples of a surgical video analysis system for segmenting surgical videos of a given surgical procedure into shorter video segments and labeling/tagging these video segments with multiple categories of machine learning descriptors. In one aspect, a process for processing surgical videos recorded during performed surgeries of a surgical procedure includes the steps of: receiving a diverse set of surgical videos associated with the surgical procedure; receiving a set of predefined phases for the surgical procedure and a set of machine learning descriptors identified for each predefined phase in the set of predefined phases; for each received surgical video, segmenting the surgical video into a set of video segments based on the set of predefined phases and for each segment of the surgical video of a given predefined phase, annotating the video segment with a corresponding set of machine learning descriptors for the given predefined phase.