TRISO aims to provide cardiac parameters such as EF, CO, RV/LV, IVC size, Collapse Rate and Pericardial Effusion (and more). Software assisted guidance is used to help achieve optimal positions and machine parameters for Sub Costal 4CH, SubCostal IVC, Apical 4CH and Apical 2CH. The cardiac parameters are automatically calculated and presented once the right echo view is being reached.
Triso developed end to end algorithms (deep learning based) using Convolutional Neural Network (CNNs) transformer networks based on attentional mechanisms in order to extract customized features. Confidence levels are shown for enabling the user's to validate our analysis. Triso algorithms are optimized to run real time analysis on edge devices.
About 100,000 patient clips are being used for training the AI system and for validating the outcomes. Then another set of clips are used for testing the system performance.
Sample screenshots of the TRISO system are shown below according to the order that they are achieved during the testing procedure: