rocAL#

The AMD ROCm Augmentation Library (rocAL) is designed to efficiently decode and process images and videos from a variety of storage formats and modify them through a processing graph programmable by the user. rocAL currently provides C API. For more details, go to docs page.

Supported Operations#

rocAL can be currently used to perform the following operations either with randomized or fixed parameters:

Blend	Blur (Gaussian 3x3)	Brightness	Color Temperature
ColorTwist	Contrast	Crop	Crop Mirror Normalization
CropResize	Exposure Modification	Fisheye Lens	Flip (Horizontal, Vertical and Both)
Fog	Gamma	Hue	Jitter
Lens Correction	Pixelization	Raindrops	Random Crop
Resize	Resize Crop Mirror	Rotation	Salt And Pepper Noise
Saturation	Snowflakes	Vignette	Warp Affine

Prerequisites#

Ubuntu 20.04/22.04
AMD RPP
AMD OpenVX™ and AMD OpenVX™ Extensions: VX_RPP and AMD Media
Turbo JPEG - Version 2.0 or higher
Half-precision floating-point library - Version 1.12.0 or higher
Google Protobuf - Version 3.11.1 or higher
LMBD Library
RapidJSON
PyBind11
HIP
OpenMP
C++ Version 17

Build instructions#

rocAL builds and installs as part of the MIVisonX toolkit. The rocAL dependencies are installed with MIVisionX-setup.py. To manually install rocAL dependencies, follow below instructions.

Turbo JPEG installation#

Turbo JPEG library is a SIMD optimized library which currently rocAL uses to decode input JPEG images. It needs to be built from the source and installed in the default path for libraries and include headers. You can follow the instruction below to download the source, build and install it. Note: install nasm package

sudo apt-get install nasm

Note: You need wget package to download the tar file.

 sudo apt-get install wget

git clone -b 2.0.6.1 https://github.com/rrawther/libjpeg-turbo.git
cd libjpeg-turbo && mkdir build && cd build
cmake -DCMAKE_INSTALL_PREFIX=/usr \
      -DCMAKE_BUILD_TYPE=RELEASE  \
      -DENABLE_STATIC=FALSE       \
      -DCMAKE_INSTALL_DOCDIR=/usr/share/doc/libjpeg-turbo-2.0.3 \
      -DCMAKE_INSTALL_DEFAULT_LIBDIR=lib  \
      ..
make -j$nproc && sudo make install

LMDB installation#

sudo apt-get install libgflags-dev libgoogle-glog-dev liblmdb-dev

RapidJSON installation#

git clone https://github.com/Tencent/rapidjson.git
cd rapidjson && mkdir build && cd build
cmake ../ && make -j$nproc && sudo make install

PyBind11 installation#

pip install pytest==3.1 
git clone -b  https://github.com/pybind/pybind11 
cd pybind11 && mkdir build && cd build
cmake -DDOWNLOAD_CATCH=ON -DDOWNLOAD_EIGEN=ON ../ && make -j4 && sudo make install

Sample and test applications#

Image augmentation application demonstrates how rocAL’s C API can be used to load jpeg images from the disk, decode them and augment the loaded images with a variety of modifications.
Augmentation unit tests can be used to test rocAL’s API individually.

rocAL

Contents