Skip to content

Latest commit

 

History

History

batchedLabelMarkersAndCompression

README.md

Batched Image Connected Region Label Markers and Compression using NPP+

Description

This project demonstrates Labeling and Compression of Image Regions using the Nvidia Performance Primitives (NPP+) library. It efficiently labels connected regions in images and compresses the labels for optimized processing. This is especially useful for processing large batches of images in GPU-accelerated environments.

Key Concepts

  • Image Region Labeling: Identifies connected regions in an image and assigns labels to each region.
  • Label Compression: Compresses labeled regions to optimize memory usage and performance, reducing redundancy in the labeled regions.

Supported SM Architectures

This example is optimized for the following Nvidia GPU architectures:

Supported Operating Systems

  • Linux
  • Windows

Supported CPU Architecture

  • x86_64

Getting Started with NPP+

To use this example, first download and install the latest version of NPP+ on your Nvidia GPU-equipped machine:

NPP+ APIs Involved

This example leverages key functions from the NPP+ library for efficient image processing on Nvidia GPUs.

Architecture

This project includes the following components:

  • Labeling Connected Image Regions: Labels all connected regions in an image.
  • Compression of Labeled Regions: Compresses the labeled regions for efficient memory usage and faster processing.

Prerequisites

Before building and running the example, ensure that your system meets the following requirements:

Building the Project

Build command on Linux

$ mkdir build
$ cd build
$ cmake .. -DNPP_PLUS_PATH=/usr/lib/x86_64-linux-gnu 
$ make

Build command on Windows

$ mkdir build
$ cd build
$ cmake -DCMAKE_GENERATOR_PLATFORM=x64 ..
$ Open findContour.sln project in Visual Studio 15 2017 and build

Example Command

./batchedLabelMarkersAndCompression -h

Usage: ./batchedLabelMarkersAndCompression [-b number-of-batch]
Parameters: 
	number-of-batch	:	Use number of batch to process [default 5]

Sample Output:

$ ./batchedLabelMarkersAndCompression -b 5

./batchedLabelMarkersAndCompression

NPP Library Version 0.9.0 CUDA Driver Version: 12.7 CUDA Runtime Version: 12.6

Input file load succeeded. Lena_CompressedMarkerLabelsUF_8Way_512x512_32u succeeded, compressed label count is 572. Input file load succeeded. CT_Skull_CompressedMarkerLabelsUF_8Way_512x512_32u succeeded, compressed label count is 434. Input file load succeeded. PCB_METAL_CompressedMarkerLabelsUF_8Way_509x335_32u succeeded, compressed label count is 3733. Input file load succeeded. PCB2_CompressedMarkerLabelsUF_8Way_1024x683_32u succeeded, compressed label count is 1272. Input file load succeeded. PCB_CompressedMarkerLabelsUF_8Way_1280x720_32u succeeded, compressed label count is 1468.