Reproducible Label Reviews
Home Page: https://rlpr.github.io
This repository hosts the submissions to the Reproducible Label proposed by the RRPR workshop. It is oly managed through the Issues module.
You can find:
To submit your code, fill this template.
Give the main steps to reproduce the figure/tab
Preparation:
- Unzip Soft_Labeled_Self_Learning.zip
Tab. 2:
1. Load `Nearest Mean/Banana_Experiment_NM.mat`
2. Run `NearestMeanSS.m`
Tab. 3:
1. Move `Nearest Mean/Gauss2_LS.mat` to `Nearest Mean/Gauss2_NM.mat`
2. Load `Nearest Mean/All_Gauss_NM.mat`
3. Run `NearestMeanSS.m`
Tab. 4:
1. Load `Least Squared/All_Gauss_LS.mat`
2. Run `LeastSquaresSS.m`
Tab. 5:
1. Load `Nearest Mean/All_Sets_NM.mat`
2. Run `NearestMeanSS.m`
Tab. 6:
1. Load `Least Squared/All_Sets_LS.mat`
2. Run `LeastSquaresSS.m`
Nearest Mean/Gauss2_NM.mat doesn't exist: second line of Tab. 3 cannot be tested, tests are executed with Nearest Mean/Gauss2_LS.mat
Conference: ICPR 2016
Source paper: paper S03
Papers title: Kernel Hierarchica PCA for person re-identification
GitHub repo: https://github.com/pratesufop/ReID_framework
Submission date: 26/11/2016
Platform: Ubuntu 17.10
Language: Matlab R2017b
Results to be reproduced: (indicates reference to figure/tab): table 1 and table 2
Reproduction main step instructions:
VIPER dataset:
Download it on http://users.soe.ucsc.edu/~manduchi/VIPeR.v1.0.zip
Unzip and cp cam_x folders (with x equals to a and b) into ./datasets/viper/camX (with X in uppercase)
PRID450s dataset:
Download it on https://www.tugraz.at/institute/icg/research/team-bischof/lrs/downloads/prid450s/
Unzip and cp cam_x folders (with x equals to a and b) into ./datasets/prid450s/camX (with X in uppercase)
Preparation:
In file ReID_framework.m:
lines 9,10,11, replace \ by / in all paths
line 14, replace prid450S by prid450s
line 37, uncomment ICPR2016 instruction
lines 39 to 43, comment the last if ... else if ... end
before instruction ICPR2016 line 37, add filename = ['ICPR2016_' dataset]
In file ICPR2016.m, replace \ by / in all strings corresponding to a path
rename folder ./Graphics/prid450S into ./Graphics/prid450s
Execution for VIPeR dataset:
In file ReID_framework.m
check that line 13 dataset = 'viper'; is UNcommented and the line 14 dataset='prid450s'; is commented
insert filename = 'ICPR2016_viper' before line 37
Run ReID_framework.m
Execution for PRID450s dataset:
In file ReID_framework.m:
check that line 13 dataset = 'viper'; is commented and the line 14 dataset='prid450s'; is UNcommented
insert filename = 'ICPR2016_prid450s' before line 37
Run ReID_framework.m
Execution without errors
Results are obtained
Results correspond to the paper's figures/tables
Execution without errors:
Results are obtained:
Results correspond to the paper's figures/tables:
Run main.m Installations steps given here https://github.com/shidahe/semidense-lines/blob/master/README.md
Reproductions of figures:
Fig 4: ...
Fig 5: ...
Fig 6: ...
Table 1: ...
Table 2: ....
Conference: ICPR 2018 & RRPR18 (id 09)
Papers title: Scalable Spectral Clustering with Cosine Similarity
Source paper: paper S11
GitHub repo: https://github.com/glsjsu/rprr2018
Submission date: 5/08/2018
Platform:
Language: MATLAB
Results to be reproduced:
Reproduction main step instructions: (from [Readme]. (https://github.com/glsjsu/rprr2018/blob/master/README.md) file)
To reproduce all the results of the paper, just do the following:
download the ssc-cosine.zip file and uncompress it
download all the .mat files and store them in a subfolder called Data
run script_all.m from the parent folder in MATLAB.
Structure of the package
Main function: ssc-cosine.m
Scripts used to reproduce the individual results reported in the ICPR18 paper:
script_20news_processing: This script processes the raw 20newsgroups data (Matlab bydate version) downloadable from http://qwone.com/~jason/20Newsgroups/ (executing this script is optional as the processed data has been provided).
script_20news_results.m: Table I
script_20news_insights.m: Figure 2
script_20news_alpha.m: Figure 3
script_tdt2_top30_results.m: Figure 4
script_digits_results.m: Table II
Scripts used to reproduce the results reported in the short paper:
script_20news_alpha_scalable3.m: Figure 1
script_tdt2_top30_DMt.m: Figure 2
Required external functions:
The kmeans.m function, available through the Statistics and Machine Learning Toolbox, is needed by the main function ssc-cosine.m. If that toolbox is not available in the computer, then one may use instead a substitute kmeans implementation, such as the litekmeans.m function available at http://www.cad.zju.edu.cn/home/dengcai/Data/Clustering.html.
The bestMap.m function, available also on the above webpage, is needed by the scripts for finding the best match between the ground-truth labels and the group labels obtained by the function ssc-cosine.m, in order to compute the clustering accuracy.
For your convenience, the litekmeans.m and bestMap.m functions have been included in this repository.
Execution without errors
Results are obtained
Results correspond to the paper's figures/tables
Execution without errors:
Results are obtained:
Results correspond to the paper's figures/tables:
The execution of the commands did not pose any particular problem and the output was produced as expected.
Results relative to accuracies (Table III) are not exactly the same as the ones presented in the paper. However, as stated by the authors, "as the algorithms are non-deterministic approaches, it's possible to get different classification results within standard deviation ranges". The ranking of the different methods is however consistent with the results presented in the paper.
Results relative to execution time could however not really be related to the ones presented in Table IV.
ICPR: A novel pattern-based edit distance for automatic log parsing - RRPR: A novel pattern-based edit distance for automatic log parsing: Implementation and reproducibility notesC++ and Python 3 - jupyter notebookfull_experiments.ipynbConference: ICPR 2016
Source paper: paper S02
Papers title: Building Facade Recognition from Aerial Images using Delaunay Triangulation Induced Feature Perceptual Grouping
GitHub repo: https://github.com/NathanUA/BuildingFacadeGrouping
Submission date: 26/11/2016
Platform: MacOS 10.12.1 / Ubuntu 17.10
Language: Matlab R2016a / Matlab R2017bĂŠ
Results to be reproduced: (indicates reference to figure/tab): Fig 4, Fig 5, Fig 6
Reproduction main step instructions:
First row: original data
Data: image1.png, image2.png, image3.png, image4.png
Second row:
Code: ICPR_2016_XuebinQin_Demo.zip
Run: go to unzipped ICPR_2016_XuebinQin_Demo.zip and run SURFFeaturesDetection.m
Third row: ICPR_2016_XuebinQin_Demo.zip.
Code: go to unzipped ICPR_2016_XuebinQin_Demo.zip and run DEMO.m. The result shown in
figure(2) is that, just using different drawn marker compared with the figure in our paper.
Fourth row:
Code: PAMI09Win7Matlab201264bit.izp. It was downloaded from http://vision.cse.psu.edu/data/data.shtml (Lattice Detection Code for MATLAB) (tested on Win7 64bit and Matlab 2014a )
Run: to run it just read the readme file then run DEMO1.m and I have already put images in the root file.
Notes: The algorithm will generate different results for each time you run because it depends on an RANSAC like sample idea.
Tested platform: MacOS 10.12.1 / Ubuntu 17.10
Reviewer comment:
The result were correctly reproduced from the information given by the authors. Here some comments:
Fig 4:
Fig 5:
Fig 6:
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
đ Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
Django
The Web framework for perfectionists with deadlines.
Laravel
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. đđđ
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
Microsoft
Open source projects and samples from Microsoft.
Google
Google â¤ď¸ Open Source for everyone.
Alibaba
Alibaba Open Source for everyone
D3
Data-Driven Documents codes.
Tencent
China tencent open source team.