Release Notes

VPI-4.1.3

VPI-4.1.3 is the latest update in the VPI-4.1 branch. These release notes cover the user-visible changes from VPI-4.0.7 through VPI-4.1.3.

For previous release notes refer to Previous Releases

Release Highlights

• Added support for Jetson Orin platforms in the VPI-4.1 branch.
• Added vpiSubmitCUDAHostFunction and VPICUDAHostFunction for submitting custom CUDA work in VPI stream order.
• Added VPI_IMAGE_BUFFER_NVSCIBUF support for wrapping NvSciBuf-backed VPI images.
• Added VPI_APRILTAG_36H9 support for AprilTag detection.
• Added sample 21 showing how to wrap CUDA device memory, host memory and NvBufSurface into VPI images.
• Added sample 22 showing how to submit custom CUDA and NPP work through vpiSubmitCUDAHostFunction.
• Added sample 23 showing how to extract host, CUDA, NvBufSurface and NvSciBuf views from a VPIImage through vpiImageLockData.
• Tegra image interop: Added vpiImageLockData support for exporting VPI_IMAGE_BUFFER_NVSCIBUF and VPI_IMAGE_BUFFER_NVBUFFER handles from VPI images.
• PVA Backend: Added a broad set of pva-solutions-backed implementations, including Background Subtractor, Bilateral Filter, Brute Force Matcher, Canny Edge Detector, Convert Image Format, Equalize Histogram, Histogram Even, Image Flip, Image Stats, Min/Max Location, Perspective Warp and Remap.
• PVA Backend: Added block-linear image support for selected pva-solutions-backed algorithms and formats on Tegra platforms.
• Compatibility with CUDA 13.2
• Added benchmark coverage for the new pva-solutions-backed implementations.

Other Updates

• On Orin platforms, requesting VPI_IMAGE_BUFFER_NVBUFFER through vpiImageLockData is still not supported. Use VPI_IMAGE_BUFFER_CUDA_PITCH_LINEAR or VPI_IMAGE_BUFFER_NVSCIBUF instead.
• Added VPI_SWIZZLE_Z000, VPI_SWIZZLE_0Z00, VPI_SWIZZLE_00Z0 and VPI_SWIZZLE_000Z for Z-channel extraction.

Resolved Issues

• StereoDisparity: Block-linear NV12 input formats are now supported by OFA and OFA+PVA+VIC backends, including VPI_IMAGE_FORMAT_NV12_BL, VPI_IMAGE_FORMAT_NV12_BL16, VPI_IMAGE_FORMAT_NV12_ER_BL and VPI_IMAGE_FORMAT_NV12_ER_BL16.
• Sample 18 ORB: C++ and Python samples now accept the PVA backend. PVA ORB support uses FAST corner scoring.

Known Issues

• Dense Optical Flow: There is no discernible difference in quality from low, medium to high.
• Background Subtractor: Initial frames may produce unstable background model.
• PyTorch: PyTorch/CUDA interoperability requires a PyTorch module compatible with CUDA Toolkit 13.2.

Notices

Disclaimer

ALL NVIDIA DESIGN SPECIFICATIONS, REFERENCE BOARDS, FILES, DRAWINGS, DIAGNOSTICS, LISTS, AND OTHER DOCUMENTS (TOGETHER AND SEPARATELY, "MATERIALS") ARE BEING PROVIDED "AS IS." NVIDIA MAKES NO WARRANTIES, EXPRESSED, IMPLIED, STATUTORY, OR OTHERWISE WITH RESPECT TO THE MATERIALS, AND EXPRESSLY DISCLAIMS ALL IMPLIED WARRANTIES OF NONINFRINGEMENT, MERCHANTABILITY, AND FITNESS FOR A PARTICULAR PURPOSE.

Information furnished is believed to be accurate and reliable. However, NVIDIA Corporation assumes no responsibility for the consequences of use of such information or for any infringement of patents or other rights of third parties that may result from its use. No license is granted by implication of otherwise under any patent rights of NVIDIA Corporation. Specifications mentioned in this publication are subject to change without notice. This publication supersedes and replaces all other information previously supplied. NVIDIA Corporation products are not authorized as critical components in life support devices or systems without express written approval of NVIDIA Corporation.

Trademarks

NVIDIA and the NVIDIA logo are trademarks or registered trademarks of NVIDIA Corporation in the U.S. and other countries. Other company and product names may be trademarks of the respective companies with which they are associated.

Copyright

© 2019-2026 NVIDIA Corporation. All rights reserved.