Automated Stool Classification via Mobile Image Processing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Clinicians face challenges in objectively assessing therapy progression and prognosis due to a lack of continuous data collection, leading to inappropriate decision-making in healthcare, as in-hospital assessments are infrequent and out of context with patients' day-to-day activities, and random sampling issues mask changes over time.
Innovation Solution
A system and method for automated patient stool monitoring using mobile devices to collect and classify gastrointestinal data through digital image processing, including color editing and machine learning algorithms, enabling continuous and long-term patient monitoring and compliance with clinical protocols.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If in-hospital or in-clinic assessments are used for patient monitoring, then clinicians can obtain medical data, but the data collection is too infrequent and out of context with patients' day-to-day activities, leading to random sampling issues that mask changes over time
Solution Approach 1:
The patent replaces manual, in-clinic assessment mechanisms with an automated mobile device-based system that captures stool images and automatically classifies them using machine learning algorithms. This substitution enables continuous monitoring in the patient's natural environment, eliminating the infrequency and contextual disconnect of traditional in-hospital assessments while maintaining or improving measurement precision through automated Bristol Stool Chart classifications.
2Productivity
If automated stool classification systems are implemented, then continuous data collection improves, but the device complexity and processing requirements increase
Solution Approach 1:
The system implements self-service through automated machine learning algorithms that perform Bristol Stool Chart classifications without requiring clinician intervention. The mobile device automatically captures images, processes them through pre-trained models, and generates classifications, enabling continuous data collection while minimizing the need for complex manual processing infrastructure.
Solution Approach 2:
The patent applies preliminary action by pre-training machine learning models with standardized Bristol Stool Chart classification criteria before deployment. This preliminary preparation allows the system to perform accurate classifications during actual use without requiring complex real-time processing or expert intervention, thereby improving productivity while managing device complexity.
3Ease of manufacture
If manual stool assessment protocols are used, then implementation is simple, but patient compliance with clinical protocols deteriorates due to the burden of frequent in-person visits
Solution Approach 1:
The patent replaces manual assessment protocols with an automated mobile device system that performs stool classification through machine learning. This substitution maintains simplicity by using off-the-shelf mobile devices with pre-loaded algorithms, while dramatically improving patient compliance by eliminating the need for frequent in-person clinic visits and allowing patients to monitor themselves in their own environments.
Data Source
AI summary
A method of data collection of stool data via a mobile device operable to enable monitoring of gastrointestinal function. A related method of long-term monitoring of patient gastrointestinal function, using one or more signal processing tools (e.g. machine learning algorithms) for automatically interpreting patient stool data, including real-time patient-assessments, in order to detect an adverse clinical event from patient stool data. A system for facilitating real-time monitoring the gastrointestinal function, the system comprising: a camera on a mobile device, a user interface that facilitates self-monitoring of stool characteristics, so as to create health-monitoring data; mobile device storage, server storage, and remote storage (with at least one communication link between them) for storing some or all of the health-monitoring data; and a processor for interpreting such health-monitoring data for clinical or other health-monitoring application.