图像矩阵matlab代码-OptimalTransportNIPS17:NIPS2017焦点文件代码:JasonAltschul

  • w8_209159
    了解作者
  • 1.6MB
    文件大小
  • zip
    文件格式
  • 0
    收藏次数
  • VIP专享
    资源类型
  • 0
    下载次数
  • 2022-06-09 02:52
    上传日期
图像矩阵matlab代码
OptimalTransportNIPS17-master.zip
  • OptimalTransportNIPS17-master
  • plot_ot_vs_gcpb.m
    2.8KB
  • plot_varyingforeground.m
    4.3KB
  • plot_ot_smalleta.m
    4.8KB
  • plot_matrixscaling.m
    8KB
  • input_generation
  • synthetic_img_input.m
    1.1KB
  • .DS_Store
    6KB
  • ot_input_between_imgs.m
    599B
  • mnist
  • loadMNISTImages.m
    811B
  • view_mnist_image.m
    315B
  • t10k-labels-idx1-ubyte
    9.8KB
  • loadMNISTLabels.m
    516B
  • t10k-images-idx3-ubyte
    7.5MB
  • algorithms
  • frobinnerproduct.m
    62B
  • gcpb_compute_ot.m
    802B
  • round_transpoly.m
    492B
  • greenkhorn.m
    3.6KB
  • computeot_lp.m
    559B
  • gcpb_ot.m
    1.3KB
  • sinkhorn.m
    1.5KB
  • readme.txt
    1.9KB
内容介绍
########################################################## # Readme file for MATLAB code for NIPS 2017 paper: # "Near-linear time approximation algorithms for optimal transport via Sinkhorn iteration" # -- Jason Altschuler, Jonathan Weed, Philippe Rigollet ########################################################## Each MATLAB file has documentation inside of it. Here's a high-level overview. -- NB: our focus writing this code was clarity over speed. Much further optimization can certainly be done. -- NB: all references to our "paper" are to the revised NIPS 2017 version, i.e. "https://papers.nips.cc/paper/6792-near-linear-time-approximation-algorithms -for-optimal-transport-via-sinkhorn-iteration" ########################################################## ALGORITHMS: -- sinkhorn.m: Implementation of classical Sinkhorn algorithm for matrix scaling. -- greenkhorn.m: Implementation of our new Greenkhorn algorithm for matrix scalng. -- compute_ot_lp.m: Compute optimal transport directly using a MATLAB linear program solver. -- round_transpoly.m: Implementation of our algorithm for rounding to transport polytope. ########################################################## PLOTTING SCRIPTS: -- plot_matrixscaling.m: (Figure 2 in paper) Compares Sinkhorn vs Greenkhorn for matrix-scaling. -- plot_varyingforeground.m: (LHS of Figure 3 in paper) Compares Sinkhorn vs Greenkhorn for matrix-scaling, when amount of "salient" data is varied. -- plot_ot_smalleta.m: (RHS of Figure 3 in paper) Compares Sinkhorn vs Greenkhorn-based OT with small regularization parameter eta. -- plot_ot_gcpb.m: (Figure 1 in supplement of paper) Compares Greenkhorn vs algorithm from [GCPB '16] ########################################################## INPUT-GENERATION: all in 'input_generation/' sub-directory. Contains code creating OT instances out of images (either synthetically generated or MNIST).
评论
    相关推荐