A11 Visual intelligent system for tele-medicine endoscopy applications (VISTA)

Battlefield assessment of injured personnel during LSCO (Large Scale Combat Operations), including those with functional gastrointestinal disorders often triggered by high stress in extreme environments, requires rapidly deployable diagnostic solutions. Wireless Capsule Endoscopy (WCE) offers a minimally invasive alternative to traditional endoscopy under such conditions. Our solution, the Visual Intelligent System for Tele-medicine Endoscopy Applications (VISTA), leverages artificial intelligence (AI) to act as a ‘first reader’ of WCE data—flagging relevant diagnostic information and significantly reducing review time. This enables military medical personnel to focus on critical care, enhancing both diagnostic accuracy and operational readiness within the Advanced First Responder Capabilities domain. Wireless Capsule Endoscopy Systems (WCES) involve ingesting a pill-sized capsule equipped with a high-resolution camera. WCES provides major advantages over traditional endoscopy, especially for monitoring the health of personnel in remote or hazardous environments. It can also help differentiate between conditions with overlapping gastrointestinal symptoms—such as radiation exposure, which is a known risk in specific military contexts and may present with nausea, vomiting, or internal bleeding. Early identification of such conditions via WCE enables timely triage and evacuation to off-field hospitals, ensuring that soldiers remain operationally effective. Although capsule endoscopy is simple to perform, the review process is highly demanding: each exam produces approximately 50,000 images, and abnormalities may appear in only a handful of frames. In scenarios where entire units are periodically screened, the volume of data becomes overwhelming and makes traditional review infeasible in the field. VISTA addresses this bottleneck by combining AI and explainable AI (XAI) to build an integrated system. This includes a Computer-Aided Detection (CADe) module for gastrointestinal pathologies, interactive tools for human reviewers, and automated patient report generation based on validated findings. The system is being developed as a cross-vendor solution deployable either in the tactical cloud or on local edge devices, enabling Tactical Field Care (TFC) and Enhanced Field Care (EFC). Data dissemination is supported via secured 5G/NTN (Non-Terrestrial Networks) and Mobile Ad Hoc Network (MANET) architectures, both airborne and ground-based. VISTA’s integrated approach is based on three pillars: 1) Edge computing capabilities, 2) Real-time telemedicine data dissemination, and 3) Validated XAI architectures. Together, these components empower VISTA’s CADe ‘first reader’ to deliver high-sensitivity, low-latency triage support. This reduces the review burden on human readers and enables faster, more informed decision-making—particularly in LSCO-deployed scenarios. Modern AI algorithms, including vision transformers and convolutional neural networks, form the backbone of the system’s diagnostic accuracy and scalability.