Radeon Performance Primitives (RPP) library is a comprehensive high-performance computer vision library for AMD (CPU and GPU) with HIP and OpenCL back-ends.
RPP is developed for Linux operating system.
- Ubuntu
16.04/18.04/20.04 - ROCm supported hardware
- ROCm
- CMake Version
3.5or higher - Boost
- Clang: minimum version required: 6.0
- IEEE 754-based half-precision floating-point library - half.hpp
NOTE:
- Boost
sudo apt --yes install libboost-all-dev
- CLang
sudo apt --yes install clang
- half.hpp
wget https://sourceforge.net/projects/half/files/half/1.12.0/half-1.12.0.zip
unzip half-1.12.0.zip -d half-files
sudo cp half-files/include/half.hpp /usr/local/include/
- Brightness modification
- Contrast modification
- Hue modification
- Saturation modification
- Color temperature modification
- Vignette effect
- Gamma Correction
- Histogram Balance
- Exposure modifications
- Foggy
- Rainy
- Snowy
- RandomShadow
- Rotate
- Warp-affine
- Flip (horizontally or vertically)
- Fish Eye Effect
- Lens correction
- Resize
- RandomResizeCrop
- Blending images
- Adding Occlusion
- Pixilation
- Adding Noise
- Blurring
- Adding Jitter
- RandomCropLetterBox
- Absolute Difference
- Accumulate
- Accumulate Squared
- Accumulate Weighted
- Arithmetic Addition
- Arithmetic Subtraction
- Bilateral Filter
- Bitwise AND
- Bitwise EXCLUSIVE OR
- Bitwise INCLUSIVE OR
- Bitwise NOT
- Box Filter
- Canny Edge Detector
- Channel Combine
- Channel Extract
- Control Flow
- Convert Bit Depth
- Custom Convolution
- Data Object Copy
- Dilate Image
- Equalize Histogram
- Erode Image
- Fast Corners
- Gaussian Filter
- Gaussian Image Pyramid
- Harris Corners
- Histogram
- Integral Image
- LBP
- Laplacian Image Pyramid
- Magnitude
- MatchTemplate
- Max
- Mean and Standard Deviation
- Median Filter
- Min
- Min, Max Location
- Non-Linear Filter
- Non-Maxima Suppression
- Phase
- Pixel-wise Multiplication
- Reconstruction from a Laplacian Image Pyramid
- Remap
- Scale Image
- Sobel 3x3
- TableLookup
- Tensor Add
- Tensor Convert Bit-Depth
- Tensor Matrix Multiply
- Tensor Multiply
- Tensor Subtract
- Tensor TableLookUp
- Tensor Transpose
- Thresholding
- Warp Affine
- Warp Perspective
- Support for 3C(RGB) and 1C(Grayscale) images
- Planar and Packed
- Host and GPU
- Batch Processing with 26 variations
- ROI variations
- Padded Variations
$ git clone https://github.com/GPUOpen-ProfessionalCompute-Libraries/rpp.git
$ cd rpp
$ mkdir build
$ cd build
$ cmake -DBACKEND=OCL .. #for OCL and HOST
or
$ cmake -DBACKEND=HIP -DCOMPILE=STATIC #for HIPSTATIC
or
$ cmake -DBACKEND=HIP -DCOMPILE=HSACOO #for HIPHSACOO
or
$ cmake -DBACKEND=HIP -DCOMPILE=HIPRTC #for HIPRTC
$ make -j4
$ sudo make install
Extended RPP support as a functionality through OpenVX MIVisionX (Find build instructions and build the amd_rpp library)
err = clGetPlatformIDs(1, &platform_id, NULL);
err = clGetDeviceIDs(platform_id, CL_DEVICE_TYPE_GPU, 1, &device_id, NULL);
theContext = clCreateContext(0, 1, &device_id, NULL, NULL, &err);
theQueue = clCreateCommandQueue(theContext, device_id, 0, &err);
d_a = clCreateBuffer(theContext, CL_MEM_READ_ONLY, bytes, NULL, NULL);
d_c = clCreateBuffer(theContext, CL_MEM_WRITE_ONLY, bytes, NULL, NULL);
err = clEnqueueWriteBuffer(theQueue, d_a, CL_TRUE, 0, bytes, h_a, 0, NULL, NULL);
cl_mem d_f;
d_f = clCreateBuffer(theContext, CL_MEM_READ_ONLY, f_bytes, NULL, NULL);
err = clEnqueueWriteBuffer(theQueue, d_f, CL_TRUE, 0, f_bytes, h_f, 0, NULL, NULL)
Rpp32f alpha=2;
Rpp32s beta=1;
RppiSize srcSize;
srcSize.height=height;
srcSize.width=width;
rppi_brighten_8u_pln1_gpu( d_a, srcSize, d_c, alpha, beta, theQueue);//device side API call
rppHandle_t handle;
rppCreateWithBatchSize(&handle, noOfImages);
rppi_resize_u8_pkd3_batchDD_host(input, srcSize, output, dstSize, noOfImages, handle);
Rpp32f alpha=2;
Rpp32s beta=1;
RppiSize srcSize;
srcSize.height=height;
srcSize.width=width;
rppi_brighten_8u_pln1_gpu( d_a, srcSize, d_c, alpha, beta, theQueue);//device side API call
hipMalloc(&d_input, ioBufferSize * sizeof(Rpp8u));
hipMalloc(&d_output, ioBufferSize * sizeof(Rpp8u));
check_hip_error();
hipMemcpy(d_input, input, ioBufferSize * sizeof(Rpp8u), hipMemcpyHostToDevice);
check_hip_error();
Rpp32f alpha=2;
Rpp32s beta=1;
RppiSize srcSize;
srcSize.height=height;
srcSize.width=width;
rppi_brightness_u8_pkd3_gpu(d_input, srcSize[0], d_output, alpha, beta, handle); //device side API call
RPP with GDF(uses OpenVX) code snippet
# specify input source for input image and request for displaying input and output images
read input ../images/face.jpg
view input inputWindow
view output brightnessWindow
# import RPP library
import vx_rpp
# create input and output images
data input = image:480,360,RGB2
data output = image:480,360,U008
# compute luma image channel from input RGB image
data yuv = image-virtual:0,0,IYUV
data luma = image-virtual:0,0,U008
node org.khronos.openvx.color_convert input yuv
node org.khronos.openvx.channel_extract yuv !CHANNEL_Y luma
# compute brightness and contrast in luma image using Brightness function
data alpha = scalar:FLOAT32,1.0 #contrast control
data beta = scalar:INT32,30 #brightness control
node org.rpp.Brightness luma output alpha beta