Medical Data Mining TRIZ Cases
- Individually-Pruned Neural Network for Cuffless Blood Pressure Estimation
An individually-pruned neural network estimates blood pressure from seismocardiogram data using separability-ranked filters.
- Heart Graphic Display System for Arrhythmia Source Localization
A heart graphic display system generates intra-cardiogram similarity and source location graphics to visualize arrhythmia patterns.
- Wireless Gastrointestinal EMG Patches for Ambulatory Monitoring
Wearable wireless patches with distributed electrode arrays capture gastrointestinal electrical activity.
- Glucose Sensor Variability Reduction via Machine Learning Correction
Machine learning models estimate and correct glucose sensitivity using real-time sensor data to reduce variability.
- Oral Microbiota Screening Model for Esophageal Cancer Detection
A logistic regression model using specific oral microbiota markers to detect esophageal cancer.
- Voice-Based Trauma Screening Using Deep Learning
Converting voice signals to spectrograms enables non-contact trauma detection, bypassing patient rejection barriers inherent in direct clinical assessments.
- Body Composition Index Calculation for Fair User Performance Comparison
A computing device generates a normalized index value by dividing actual body composition change time against a reference group period.
- Generalized Operational Perceptrons for Complex Neural Network Learning
Generalized Operational Perceptrons encapsulate multiple operators within each neuron to enable adaptive network structures.
- Probabilistic Medical Profile Generation Using Age-Correlated Patient Data
A computer system generates future probabilistic medical profiles by aggregating age-correlated characteristics from multiple patients.
- Glaucoma Diagnostic Data Harmonization via Common Scale Conversion
A combinatorial analysis system transforms structural and functional test measurements into a common distribution for unified assessment.
- Smart Vital Device Health Monitoring via ML and Sensors
Smart vital device detects health conditions using machine learning on audio and temperature sensor data.
- Adaptive Procedure Support System for Medical Workflows
A trained function applies real-time adjustments to capture time-resolved procedure data and provide adaptive support.
- Private AI Training for Patient Assessment
A patient assessment system trains AI algorithms using computer code executed on local data systems to process sensitive information securely.
- Automated Neural Network Pre-processor for Model Accuracy
Pre-processor identifies machine learning problem types and generates optimized input datasets, reducing training time while maintaining model accuracy.
- Tensor-Based Graph Processing Model Deficiency Data Objects
A tensor-based graph processing framework generates model deficiency data objects using holistic graph links inferred by a representation machine learning model.
- Reasoning Engine Automating Ambient Data Inference and Validation
A reasoning engine acquires environment data to generate hypotheses about correlated aspects within the observed space.
- Ovarian Stimulation Algorithm for Egg Quality via FSH Control
Extending the follicular phase through dynamic FSH adjustment resolves the trade-off between egg quality and treatment duration.
- Gait Analysis System Using 3-Axis Sensor Data and Energy Density Spectra
Wearable leg sensors capture acceleration data for a computing unit to compute energy density spectra from stride characteristics.
- Personal Health Monitoring via Correlation-Based Parameter Selection
A health monitoring system selects secondary parameters correlated with primary metrics to reduce resource usage.
- Neural Network Recipe Generation for Market-Aligned Food Products
An AI platform replaces manual recipe development with automated generation that aligns product flavor and texture with current market trends.
- Medical Search System Using Prescription Data to Resolve Disease Name Variations
A medical search system acquires related disease names from prescription records to retrieve relevant electronic health information.
- NFT-Based EHR Tokenization for Interoperability and Security
A blockchain system mints non-fungible tokens for electronic health records, resolving interoperability conflicts while maintaining strict data security.
- AI-Controlled Synthetic Thyroid Valve for Hormone Homeostasis
An intelligent valve delivery mechanism adjusts synthetic thyroid hormone flow based on real-time blood sensor data.
- Contrastive Learning Sleep Stage Classification Reducing User Performance Variation
A contrastive learning sleep classification system extracts unique biosignal features to standardize model inputs across diverse users.
- Patient Cohort Response Prediction via Interactive Data Filtering
Automated system analyzes molecular data through interactive interfaces to resolve the trade-off between comprehensive analysis speed and prediction accuracy.
- AI ECG Workflow Automation for Immediate Patient Analysis
Automated AI analysis of ECG waveforms generates immediate clinical recommendations, eliminating intermediate cardiologist visits and reducing treatment time.
- Sleep Event Classification Model Finetuning for Data Scarcity
Finetune a pre-trained sleep classification model with target population data to resolve source-target mismatch and overcome training data scarcity.
- Wearable Sensor System for Craving Detection via Physiological Signals
A wearable sensor system monitors physiological signals to detect cravings in individuals recovering from addiction.
- Head-Mounted Device Health Metrics via Motion Sensor
A head-mounted device uses a motion sensor to capture pulsatile movements for extracting health metrics.
- Clinical Report Retrieval and Comparison via Semantic Timeline
Classifies and retrieves longitudinal clinical reports, then builds a model to predict semantic relationships for visual timeline construction.
- Neurological Signal Monitoring System for Seizure Prediction
A neurological signal monitoring system analyzes digital data strings to identify recurring patterns in brain activity.
- AI Sleep Apnea Screening Using Wearable Pulse Oximetry
Machine learning model analyzes pulse oximetry data from wearable sensors to classify sleep apnea risk.
- Decision Engine for Medical Image Diagnosis with Incomplete Clinical Data
A decision engine processes medical images to generate diagnostic outcomes using ensemble classifiers and Bayesian analysis.
- Medical Concept Search Engine with Bubble Graph Interface
A medical concept searching engine generates a bubble graph interface to visually represent related search criteria for intuitive navigation.
- Bayesian Signal Detection for Recurrent Feature Identification
A Bayesian framework parameterizes signals using probability density functions to determine posterior probabilities of feature hypotheses.
- Contrastive Learning for Pseudo-Severity Label Generation
An anomaly detection module produces severity scores to train models without manual labeling, improving biomarker classification accuracy.
- Machine Learning Payor Class Prediction for Medical Claims
A machine learning system predicts medical claim payor class by analyzing standardized codes.