CNN Surgical Object Counting Without Image Databases

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing systems for identifying and counting surgical objects during medical procedures are inefficient and inaccurate, particularly for new or unknown objects, and require extensive database maintenance and shape recognition algorithms that can be misled by suture tails and other debris.

Innovation Solution

A system utilizing a convolutional neural network (CNN) for object identification and counting, which captures images of used objects, classifies them with confidence scores, and reconciles the counts without relying on pre-existing databases or shape comparisons, allowing differentiation between objects and debris.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If database-image comparison methods are used for sharp identification, then known sharps can be identified, but new/unknown sharps cannot be automatically identified and additional resources are required to maintain the database

Engineering Contradiction:

ImproveAbility to identify new/unknown sharpsVSAvoidDatabase maintenance requirements

Core Design Contradiction:

Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The system performs preliminary actions by capturing images of sharps during the procedure and processing them through CNN-based identification before the procedure concludes. This allows the system to proactively identify both known and unknown sharps without requiring database updates during the procedure, resolving the contradiction between identifying new sharps and maintaining database complexity.

Inventive Principle:

Principle #10Preliminary action

2Measurement precision

If detailed shape recognition algorithms are used, then shape and curvature comparisons can be performed, but suture tails and debris can mislead the recognition process resulting in inaccurate identification

Engineering Contradiction:

ImproveSharp identification accuracyVSAvoidMisidentification due to suture tails and debris

Core Design Contradiction:

Measurement precisionVSObject-affected harmful factors

Solution Approach 1:

The patent introduces an intermediary mechanism - a CNN-based identification system - that processes sharp images without relying on traditional shape and curvature comparisons. This intermediary approach filters out misleading elements like suture tails and debris by using learned features from training data, thereby improving identification accuracy while eliminating the harmful effects of false shape recognition.

Inventive Principle:

Principle #24Intermediary (Mediator)

3Reliability

If manual counting methods are used, then object counts can be tracked, but mistakes can occur and additional surgeries may be required to remove missed objects

Engineering Contradiction:

ImproveObject accounting accuracyVSAvoidTime for additional surgeries

Core Design Contradiction:

ReliabilityVSLoss of time

Solution Approach 1:

The patent replaces the mechanical manual counting system with an automated CNN-based identification and counting system. The automated system processes images of sharps and generates accurate counts without human intervention, eliminating mistakes associated with manual counting. This substitution ensures reliable object accounting and prevents the need for additional surgeries, thereby improving reliability while avoiding time loss.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Data Source

PatentUS12573188B2Object counting system using convolutional neural network for medical procedures

Publication Date: 2026.03.10 THE CLEVELAND CLINIC FOUND

AI summary

The present disclosure relates to a system and method for recognizing objects used in a medical procedure using a convolutional neural network. No database of image information for such objects is used or required. Rather, the neural network is trained to recognize the objects, and does not require any such image database. The system is able to reconcile the recognized objects against a ‘counted-in’ list of objects for the procedure, to ensure that all such objects are accounted for prior to closing the procedure.