The AI sensor processing SDK for low latency streaming workflows

ai cloud cuda edge embedded gpu hpc jetson medical-devices nvidia sensor streaming
3 Open Issues Need Help Last updated: Jun 30, 2025

Open Issues Need Help

View All on GitHub

AI Summary: The task is to determine the best approach for unit testing C++ operators within the NVIDIA Holoscan SDK. The current approach involves separating the operator's implementation from its Holoscan wrapper to facilitate testing, but a confirmation of this method or a suggestion of a preferred alternative is needed.

Complexity: 4/5
help wanted

The AI sensor processing SDK for low latency streaming workflows

C++
#ai#cloud#cuda#edge#embedded#gpu#hpc#jetson#medical-devices#nvidia#sensor#streaming

AI Summary: The task involves debugging an inconsistency in the Holoscan SDK's visualization operator (HolovizOp) where geometry and image layers are rendered differently when using InputSpec.View and transformation matrices. The issue manifests as incorrect positioning of geometry elements relative to transformed images. The solution requires understanding the Holoscan SDK's coordinate systems and matrix transformations, modifying the provided example code to reproduce the bug, and then fixing the underlying issue within the HolovizOp or its interaction with InputSpec.View.

Complexity: 4/5
bug help wanted

The AI sensor processing SDK for low latency streaming workflows

C++
#ai#cloud#cuda#edge#embedded#gpu#hpc#jetson#medical-devices#nvidia#sensor#streaming

AI Summary: The task is to fix a compatibility issue in the Holoscan SDK v3.2.0-dgpu container image. The issue prevents the image from running correctly under rootless Podman due to UID/GID conflicts within a specific layer (ONNX Runtime). The solution likely involves modifying the container build process to add `--no-same-owner --no-same-permissions` flags during the untarring of the ONNX Runtime layer, ensuring proper permissions for rootless users.

Complexity: 4/5
bug help wanted

The AI sensor processing SDK for low latency streaming workflows

C++
#ai#cloud#cuda#edge#embedded#gpu#hpc#jetson#medical-devices#nvidia#sensor#streaming