Skip to content

Computational photography library. The code is parallelized by using SIMD intrinsics and multi-threading.

License

Notifications You must be signed in to change notification settings

dreamplayer-zhang/OpenCP

 
 

Repository files navigation

OpenCP

library for computational photography

Todo

** port various filters from my library.**

  • bilateral filter approxmations

    • separable
    • bilateral grid
    • constant time O(1) bilateral
    • real-time O(1) bilateral
  • add dxt thresholding for denoising

  • add super resolution demo for iterative back projection.

  • add recursive birateral filter for gray, more efficient implimentaion, line by line chatching?

  • add domain transform filter

    • add test for domain transform filter.6/11
    • add various simd implementation for RF filter.6/11
    • add SIMD V-direction ver.
    • add other implimentation of domain transform filter
  • write detail enhancement document

  • add detail enhancement class.

  • add color joint nearest filter

  • add test for stencil and parallel filtering.

  • need stereo matching/cost filter for joint filtering.

  • need sse implimmentation cvtColorBGRA2BGR,cvtColorBGR2BGRA

  • need a floating birateral filter without LUT or quantization for back projection.

add joint nearest filter 6/18
add view synthesis class 6/18
add domain transform filter of RF implementation. 6/8
add guided filter 6/5
add denoising demo5/31
add weighted birateral filter and joint birateral filter, but some implimentation is not same as non weighted version.5/31 add recursive birateral filter 5/28
add detail ehnancement 5/28
add birateral iterative back projection for debluring5/27
add iterative back projection for debluring5/26
add joint birateral filter to rect kernel implimentation. 5/26
add weighted binary range filter5/25
add joint birateral filter 5/24
add massively parallel implimentaion of birateral filter 5/21
add slowest birateral filter5/20
update type of the destination of the SLIC. mean image? mesh? 5/19

Filter

###implemented and parallelized ######filter
Gaussian IIR filter

######edge preserving filter
bilateral filter and its fast implimentations or variants
*sepalable filter
*bilateral grid
*realtime O(1) birateral filter
*joint bilateral filter
*trilateral filter
*dual bilateral filter
*weighted (joint) bilateral filter

cost volume filters *3D birateral filter
*3D guided filter

trilateral filter
non-local means filter
shiftable DXT thresholding filter
guided filter
domain transform filter
weighted mode filter
constant time median filter
joint nearest filter ######segmentation
SLIC (forked from VLFeat(http://www.vlfeat.org/). The code, which is optimized by SSE and Intel TBB, is more efficient than the VLFeat.) ######upsample joint bilateral upsample
guided upsample
hqx

implimented but not optimized

filter
recursive birateral filter constant time O(1) birateral filter
L0 Smoothing
Weighted least squre (WLS) smoothing
Gaussian KD-Tree
Permutohedral Lattice
adaptive maniforld

Example of Applications

  • denoise
  • deblur
  • up-sample/single image super resolution
  • flash/non flash photograpy
  • HDR
  • colorization
  • detail enhancement
  • stylization, abstruction
  • pencil sketche
  • up-sample for pixel art, depth map
  • removing coding noise
  • blur regeneration
  • Haze remove
  • depth map estimation/refinement
  • optical flow estimation/refinement
  • alpha matting

References

About

Computational photography library. The code is parallelized by using SIMD intrinsics and multi-threading.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • C++ 99.7%
  • C 0.3%