Holoscan operators are the building blocks of Holoscan applications — reusable processing units in HoloHub for data ingestion, transformation, visualization, AI inference, and output. Browse this Holoscan ecosystem collection to accelerate development of real-time sensor processing pipelines.
Samples live RGB frames, submits either a JPEG or one bounded MP4 clip to an OpenAI-compatible multimodal reasoner, and emits incremental or completed text.
Publishes Holoscan sensor and inference messages to Foxglove over WebSocket with optional MCAP recording.
Optional live-camera operators for Atracsys hardware, including the master source and structured-light point-cloud conversion path.
Replay and visualization operator that switches between visible, infrared, structured-light, and tracking modes while projecting tracked fiducials into Holoviz overlays.
GPU-resident Holoscan SDK Operator for receiving video frames over RoCE using DOCA GPUNetIO for GPU-side CQ polling, eliminating CPU from the critical path.
The operator runs TensorRT inference for TapNext-style architectures, where an initialization model processes the fir...
The PVA Radar Operator demonstrates RADAR signal processing offload using NVIDIA's PVA (Programmable Vision Accelerat...
The PVA Radar Operator demonstrates RADAR signal processing offload using NVIDIA's PVA (Programmable Vision Accelerat...
The PVA Radar Operator demonstrates RADAR signal processing offload using NVIDIA's PVA (Programmable Vision Accelerat...
Operator for receiving video frames over RoCE (RDMA over Converged Ethernet) directly to GPU memory. In addition to the Holoscan Sensor Bridge (v2.5) implementation, this operator provides an option to skip copying metadata from the device memory to the host memory. This is useful for GPU-resident processing pipelines.
GPU-resident operator that converts CSI (Camera Serial Interface) packed data to Bayer format for image processing pipelines.
GPU-resident operator for image processing including optical black correction, histogram calculation, and white balance.
GPU-resident operator for displaying processed image frames using cuDisp for direct GPU-to-display rendering with gamma correction.
The ST 2110 Source operator receives SMPTE ST 2110-20 uncompressed video streams over IP networks.
Operators for sending and receiving messages through a UCXX endpoint.
GStreamer bridge operators for video recording and streaming in Holoscan applications
The operator provides a modular streaming server implementation with separate upstream, downstream, and resource comp...
The VideoStreamingClientOp class implements a Holoscan operator that provides bidirectional video streaming capabilit...
A unified package containing both streaming client and server operators for real-time video communication in Holoscan...
The Holoscan ROS2 Bridge extension provides interoperability between NVIDIA Holoscan and ROS2 (Robot Operating System...
The IIO Controller provides a comprehensive set of operators for interfacing with Industrial I/O (IIO) devices in Hol...
Holoscan operator for receiving and serving video data streams using NVIDIA Streaming SDK. Provides server-side streaming capabilities with support for multiple client connections and real-time video processing.
Holoscan operator for streaming video data to remote servers using NVIDIA Streaming SDK. Provides real-time video streaming capabilities with support for various video formats and configurable streaming parameters.
A collection of specialized operators for medical imaging and surgical visualization applications, providing comprehe...
Data Distribution Service (DDS) operators for RTI Connext, enabling real-time, scalable, and reliable data exchange in distributed systems using the DDS middleware standard.
This operator reads H.264/H.265 video files and emits raw encoded frames one at a time. The frames remain in their compressed format for processing by downstream operators like nv_video_decoder.
This operator encodes a video stream into a compressed format using the NVIDIA Video Codec SDK.
This operator decodes a video stream from a compressed format using the NVIDIA Video Codec SDK.
A Holoscan operator that subscribes to messages from a ZeroMQ message queue using the PUB/SUB pattern.
A Holoscan operator that publishes messages to a ZeroMQ message queue using the PUB/SUB pattern.
A Holoscan operator that enables seamless interaction with FHIR (Fast Healthcare Interoperability Resources) services...
A Holoscan operator that processes and sanitizes FHIR medical records into a standardized, AI-friendly format while m...
This operator loads DICOM studies into memory from a folder of DICOM instance files.
This operator converts medical imaging data to STL format for 3D visualization and printing.
This operator converts a DICOM series into a volumetric image for downstream analysis in medical imaging workflows.
This operator performs inference using MONAI Bundles for medical imaging tasks.
This operator selects specific DICOM series from a set of studies for further processing in medical imaging workflows.
This operator writes segmentation results into DICOM Segmentation objects for medical imaging workflows.
This operator writes encapsulated PDF documents into DICOM format for medical imaging workflows.
This operator integrates Clara Viz visualization into medical imaging pipelines.
This operator converts medical images to PNG format for visualization or storage.
This segmentation operator uses MONAI transforms and Sliding Window Inference to segment medical images.
This operator writes DICOM Structured Report (SR) objects containing text results for medical imaging workflows.
This operator provides a base class for running inference in medical imaging pipelines.
This operator publishes medical imaging data to downstream consumers or external systems.
This operator loads NIfTI (nii/nii.gz) medical images for use in Holoscan pipelines.
A Pixelation-based Deidentification Operator In medical and sensitive imaging workflows, pixelation is a common metho...
A collection of Holoscan operators for XR (Extended Reality) integration, enabling real-time XR applications with Vul...
The AJA Source operator provides functionality to capture high-quality video streams from AJA capture cards and devices.
The FFT operator takes in a tensor of complex float data, performs an FFT, and emits the resultant tensor.
The EHR Query LLM Operators are a Holoscan operator that provides a robust interface for querying and processing Elec...
The extension provides real-time detection of April tags from different tag families.
This directory contains the Holohub gRPC plugins for Holoscan SDK, including: Please refer to the gRPC h.264 Endoscop...
The RealSense Camera Operator captures synchronized color and depth frames from Intel RealSense cameras using the Rea...
This operator subscribes to DDS topics published by the RTI Connext shapes demo.
A Holoscan operator to convert packets from the Velodyne VLP-16 Lidar sensor to a point cloud tensor format.
The extension takes the output of the source video player and the output of the operator and renders the video stream...
This directory contains two operators to enable interoperability between the CVCUDA and Holoscan tensors: and .
The operator converts 3D meshes in STL format to USD (Universal Scene Description) mesh format for integration into O...
The operator provides a way to send and receive imaging data using the OpenIGTLink library.
A collection of specialized operators for medical imaging and surgical visualization applications, providing comprehe...
A collection of specialized operators for medical imaging and surgical visualization applications, providing comprehe...
A collection of specialized operators for medical imaging and surgical visualization applications, providing comprehe...
A collection of specialized operators for medical imaging and surgical visualization applications, providing comprehe...
The operator provides video stream capture from YUAN High-Tech capture cards, supporting various video formats and co...
The Prohawk Video Processing Operator is a Holoscan SDK operator that integrates Prohawk Technology Group's video res...
The operator renders a volume using ClaraViz (https://github.com/NVIDIA/clara-viz).
The extension provides LSTM (Long-Short Term Memory) stateful inference module using TensorRT.
The extension provides a codelet that converts inference output of used in the endoscopy tool tracking pipeline to be...
This operator provides support for Emergent Vision Technologies cameras as video sources.