Difference between revisions of "CUVI Release Notes"

From CUVI Wiki
Line 10: Line 10:
**Added new '''[[Function:focusStack]]'''
**Added new '''[[Function:focusStack]]'''
**Removed '''[[Function:Invert]]'''
**Removed '''[[Function:Invert]]'''
*10x Speed up '''[[Function:underwaterFilter]]'''


==Version 1.6.1==
==Version 1.6.1==

Revision as of 14:03, 5 November 2020

Version 1.6.5 (Latest Stable)

Version 1.6.1

Version 1.6.0

Version 1.5.5

Version 1.5.2

  • Major Release Date: June 12, 2015
  • Requirements:
    • Latest CUDA drivers from NVIDIA [download]
    • CUDA capable Graphics Card [see list]
    • Windows, Linux or Mac OSX
  • Color Operations:
    • Bug fix demosiacDFPD.
    • Speed up demosiacDFPD.
  • Removed Video Stabilization module
  • Several bug fixes across the board.
  • Support for latest GPUs.

Version 1.3.1

  • Release Date: June 12, 2014
  • Requirements:
    • Latest CUDA drivers from NVIDIA [download]
    • CUDA capable Graphics Card [see list]
    • Windows, Linux or Mac OSX
  • Color Operations:
    • Added demosiacDFPD that restores an RGB image from a CFA Bayer image using DFPD algorithm.
  • Added Video Stabilization module
    • Added CuviVideoStabilizer class for transparent video stabilization.
    • Added CuviMotionEstimator class to determine the global motion vector between 2 frames.
  • Core
    • Added cuvi::showImage function to display the image.
    • Added cuvi::waitKeyPress function to keep displaying the window opened by cuvi::showImage.

Version 1.3.0

  • Major Release Date: May 05, 2014
  • Requirements:
    • Latest CUDA drivers from NVIDIA [download]
    • CUDA capable Graphics Card [see list]
    • Windows, Linux or Mac OSX
  • Core:
    • Added arithmtic and logic operators for CuviImage.
    • Added channel extraction operator [] for CuviImage.
    • Added CuviPyramid class
    • Added CuviHistogram class
    • Added CuviSpecialFilters class for creating standard predefined filters.
  • Data Exchange:
    • Added getChannel function for channel extraction.
  • Arithmetic and Logical:
    • Added different point scalar functions
    • abs -> Image absolute value
    • exp -> e raised to power image
    • absDiff -> Absolute difference between 2 images
    • sqrt -> Square root
    • cbrt -> Cube root
    • floor
    • ceil
    • round
    • Added bit shifting functions (SHL and SHR)
    • Added matrix multiplication (matMul)
    • Added matrix multiplication with transpose -> A*A' or A'*A
    • Added matrix inversion for square images.
  • Color Operations:
    • Added lookup table transformation function (LUT)
    • Merged rgba2gray into rgb2gray
    • Added color space conversions:
      • Grayscale --> RGB (gray2rgb)
      • RGB <--> YCbCr (rgb2ycbcr and ycbcr2rgb)
      • RGB <--> HSV (rgb2hsv and hsv2rgb)
      • RGB <--> Lab (rgb2Lab and Lab2rgb)
  • Image Filtering:
    • Added median filtering (medianFilter)
    • Added image enhancement algorithm (fusion)
  • Image Statistics:
    • Added histogram calculation function (calcHist)
  • Misc:
    • Replaced raw pointer/array arguments with CuviScalar class for the following functions:
      • rgb2gray
      • sum, mean, min, max, minMax, countInRange

Version 1.2.1

  • Major Release Date: April 17, 2013
  • Requirements:
    • Latest CUDA drivers from NVIDIA [download]
    • CUDA capable Graphics Card [see list]
    • Windows, Linux or Mac OSX
  • New Features/Enchancements
    • pow - Raises each pixel to the specified power.
    • rectangle - draws a rectangle on the image.
    • adjust function can now take adjust parameters separately for each channel
    • haarFwd and haarInv now also support 3 channel images.
  • Framework Enhancements
    • C Interface has been discontinued.
    • All functions in C++ interface now take CuviImage references instead of pointers.
    • Added image I/O. CuviImage can now be loaded and saved directly on the disk without the requirement of a host.
    • Added namespace "device" which contains device management functions:
      • getDeviceCount
      • getCurrentDevice
      • setCurrentDevice(const Cuvi32s id);
      • syncCurrentDevice()
      • resetCurrentDevice()
      • getDeviceProperties(const int deviceID, CuviDeviceProperties& props);
  • Features removal:
    • eigenVals
    • eigenValsVecs
    • meanStdDev
    • Color2Gray (added as an overload of rgb2gray)

Version 1.2

  • Major Release Date: May 11, 2012
  • Requirements:
    • Latest CUDA drivers from NVIDIA [download]
    • CUDA capable Graphics Card [see list]
    • Windows 7, Linux or Mac OSX
  • New Features
    • CountInRange - Counts the number of pixels within the given intensity range
    • Sum - Computes the sum of pixel values of each channel in an image
    • Max - Computes the maximum pixel value of each channel in an image
    • Min - Computes the minimum pixel value of each channel in an image
    • MinMax - Computes the minimum and maximum pixel value of each channel in an image
    • Mean - Computes the mean of each channel of the input image
    • MeanStdDev - Computes the mean and standard deviation of each channel in an image
    • EigenValsVecs - Calculates eigen values and eigen vectors at each pixel in an image
    • Flip - Mirrors an image about a horizontal or vertical axis, or both
    • ImageGradients - Calculates X and Y gradients of image
    • MinEigenVal - Calculates the minimum eigen value at each pixel in an image
    • Rotate - Rotates an image with the origin (0,0) or center
    • Transpose - Takes Transpose of the image
  • Framework Enhancements
    • Instead of a single function for all color operations and one for arithmetic operations, the new build contains individual functions like RGB2YUV and Add to improve usability
    • The functions 'split' and 'merge' are renamed to ChannelSplit and ChannelMerge
    • The function 'bayerToRGB' is renamed to Demosaic
    • The function 'imFilter' is renamed to ImageFilter and its input argument have been simplified
    • The function 'dwt' and 'idwt' have been renamed to haarFwd and haarInv respectively
    • CUVI now includes a C++ Timer class that accurately profiles your application timing
    • CuviMat is introduced which is C counterpart of the class CuviImage
  • Bug Fixes


Release notes for previous versions is not available