Use a native media pipeline for this embedded system
C++ and GStreamer replaced the proposed headless-browser processing path for direct sensor access, predictable media behavior, and integration with device-control functions.
Media, Communications and Telehealth
RemoteScan - Low-Latency Remote Diagnostic Imaging Kiosk
A native C++/GStreamer/WebRTC subsystem for embedded diagnostic devices, with negligible software-added latency and WebRTC DataChannel-based remote laser guidance.
Medical sensor integrationEmbedded LinuxGStreamerWebRTCRemote device control
The system
RemoteScan is an embedded Linux medical platform that transports live media from representative devices such as ultrasound, microscope, and digital stethoscope or phonendoscope equipment to a remote clinician station. The system combines native device integration, real-time processing and encoding, WebRTC media, an application interface, and a separate control path for patient-side positioning guidance.
The initial requested concept connected medical devices through custom UVC drivers to headless Chrome and browser WebRTC. After evaluating the embedded environment and latency requirements, OPTIME proposed a native media architecture better suited to direct device access, predictable processing, and hardware control in this particular system.
Engineering relationship
OPTIME joined an existing customer engineering team for an approximately two-year engagement involving about four engineers. The team took responsibility for the native media subsystem, medical-sensor integration, custom WebRTC path, Qt/QML and TypeScript application elements, and remote guidance control.
OPTIME completed and delivered its engineering scope. The subsequent commercial status of the overall customer product is outside OPTIME’s visibility, so this case does not claim later production deployment, certification, or clinical adoption.
Engineering constraints
- Integrate heterogeneous medical-media sources directly into an embedded Linux runtime.
- Keep software processing delay negligible so network conditions remain the primary contributor to end-to-end latency.
- Separate real-time media transport from bidirectional device-control messages.
- Map a clinician’s image coordinates into safe patient-side positioning guidance without exposing actuator internals.
- Support a native application experience using Qt/QML and higher-level TypeScript workflow components.
- Describe diagnostic media transport and clinician guidance without making diagnostic, certification, or clinical-outcome claims.
What OPTIME engineered
- Native C++ integration for ultrasound, microscope, digital stethoscope/phonendoscope, and related medical sensor inputs.
- A GStreamer pipeline for capture, media processing, encoding, and real-time handoff.
- A custom WebRTC implementation for low-latency transport to the remote clinician application.
- A WebRTC DataChannel control path carrying clinician-selected image coordinates back to the patient-side device.
- Patient-side control logic connecting coordinate selection to the laser positioning mechanism.
- Qt and QML interfaces for native device/application interaction.
- TypeScript components supporting the higher-level remote workflow.
- System integration and validation across embedded hardware, media, networking, UI, and remote-control boundaries.
Architecture
Media path - medical sensors
Ultrasound, microscope, digital stethoscope/phonendoscope, and related sources connect through native device integration.
Media path - GStreamer
The embedded pipeline captures, processes, and encodes sensor media without relying on headless Chrome as the processing layer.
Media path - custom WebRTC
The native WebRTC implementation transports the live media to the remote clinician application.
Media path - clinician viewing
The clinician receives live diagnostic media for remote observation and guidance without implying automated diagnosis.
Control path - image selection
The clinician clicks a location directly in the received video image.
Control path - coordinate mapping
The application converts the selected image point into a generalized patient-side positioning command.
Control path - WebRTC DataChannel
Coordinates travel independently from the media stream through a bidirectional DataChannel.
Control path - laser guidance
Patient-side control logic moves the laser positioning mechanism to provide visible physical placement guidance.
Key engineering decisions
Separate media and control paths
WebRTC carries the clinician-facing media, while DataChannel messages carry selected coordinates back to patient-side control logic. The two paths remain conceptually and operationally distinct.
Minimize software-added delay
The native capture, processing, encoding, and transport pipeline introduced negligible additional software latency, leaving network conditions as the primary contributor to end-to-end delay.
Keep guidance under clinician control
The system translates an authorized clinician’s selected image location into visible laser guidance; it does not infer placement, diagnose a condition, or make a clinical decision.
Architecture transition
The project moved from the proposed custom-UVC/headless-Chrome/browser-WebRTC concept to a native embedded pipeline. That choice was specific to the device integration, latency, and control requirements of this platform rather than a general judgment about browser communications.
Verified capability
The delivered engineering scope connected medical sensors to a native C++/GStreamer pipeline, custom WebRTC transport, clinician viewing, and a DataChannel-based click-to-laser guidance path. The pipeline introduced negligible additional software-processing latency.
Verified result
OPTIME completed and delivered its engineering scope. Subsequent commercial, deployment, certification, and product status remain outside OPTIME’s visibility and are not claimed.
Technology & Engineering Role
- C++
- Native device integration, media processing, encoding, WebRTC, and hardware-control components.
- Embedded Linux
- Runtime environment for the patient-side device.
- GStreamer
- Real-time medical-sensor capture, processing, and encoding pipeline.
- Custom WebRTC implementation
- Low-latency media transport to the clinician application.
- WebRTC DataChannel
- Bidirectional coordinate and device-control messages.
- Qt / QML
- Native device and application user interfaces.
- TypeScript
- Higher-level application and remote-workflow components.
- Medical sensor integration
- Ultrasound, microscope, digital stethoscope, and related source devices.