Skip to content

Commit d9d687c

Browse files
committed
update the HTML reports README
1 parent 403b2ba commit d9d687c

File tree

1 file changed

+44
-73
lines changed

1 file changed

+44
-73
lines changed
Lines changed: 44 additions & 73 deletions
Original file line numberDiff line numberDiff line change
@@ -1,130 +1,101 @@
1-
This folder contains the inputs & outputs of our algorithm, on the [MAMBO](https://gitlab.com/franck.ledoux/mambo/) dataset.
1+
# Validity-first automatic polycube labeling for CAD models - supplemental material
22

3-
Contrary to a supplemental material in the PDF format where the ordering is fixed and the 3D models are compressed into a set of 2D images,
4-
you will be able to sort/filter by any column, and visualize the results in a 3D viewer.
3+
This folder contains the inputs & outputs of our algorithm, on the [MAMBO](https://gitlab.com/franck.ledoux/mambo/) and [OctreeMeshing](https://cims.nyu.edu/gcl/papers/2019-OctreeMeshing.zip) datasets, as well as results of _PolyCut_[^polycut] and _Evocube_[^evocube].
54

6-
# Important : Allow the access to local files
5+
## Open the interactive document
76

8-
However, for security reasons, internet navigators do not allow HTML files from accessing local files
9-
([CORS error](https://developer.mozilla.org/en-US/docs/Web/Security/Same-origin_policy#file_origins)),
10-
such as the 3D models. You have 2 solutions:
7+
Our supplemental material do not require an internet connection. Rendering code is inside the `js/` subfolder and 3D models inside `glb/`.
118

12-
## 1. Temporarily change the security policy of your internet navigator
9+
By default, internet navigators do not allow both HTTP access and `file:///` access ([CORS error](https://developer.mozilla.org/en-US/docs/Web/Security/Same-origin_policy#file_origins)).
10+
If you directly open `index.html` with your internet navigator, the 3D models will not be rendered.
1311

14-
### Firefox
15-
16-
Open `about:config` and change the value of `security.fileuri.strict_origin_policy` from `true` to `false`
17-
18-
Then open `index.html`
19-
20-
### Chrome
21-
22-
Execute
12+
### Recommended: Launch a HTTP server with [Python](https://www.python.org/)
2313

2414
```bash
25-
google-chrome index.html --allow-file-access-from-files
15+
python -m http.server
2616
```
2717

28-
## 2. Launch an HTTP server
18+
Then open the printed address (usually http://0.0.0.0:8000/) with your internet navigator. Python will serve `index.html` locally and provide access to the 3D models through HTTP.
2919

30-
For example with Python:
20+
### Other means to launch a HTTP server
3121

32-
```bash
33-
python3 -m http.server
34-
```
22+
- with the [miniserve](https://github.com/svenstaro/miniserve) binary: `miniserve .`
23+
- with [PHP](https://www.php.net/): `php -S localhost:8080 -t .`
24+
- with the [http-server](https://www.npmjs.com/package/http-server) Node.js package: `http-server .`
25+
- with [Ruby](https://www.ruby-lang.org/en/) and [WEBrick](https://github.com/ruby/webrick): `ruby -run -ehttpd .`
3526

36-
Then open http://0.0.0.0:8000/ with your internet navigator
27+
### Other method: Temporarily change the security policy of your internet navigator
3728

38-
# Sankey diagram
29+
- with Firefox: open `about:config` and change the value of `security.fileuri.strict_origin_policy` from `true` to `false`, then open `index.html`
30+
- with Chrome: execute `google-chrome index.html --allow-file-access-from-files`
3931

40-
Above the interactive grid, there is a collapsed section hiding a Sankey diagram, which gives high level stats about the batch execution
41-
(for our method only, Evocube results are ignored).
42-
You can adjust the vertical position of the nodes, and hovering them gives the number of models.
32+
## Sankey diagram
4333

44-
# Column filters
34+
Above the interactive grid, there is a collapsed section hiding a Sankey diagram, which gives high level stats on our method only. You can adjust the vertical position of the nodes, and hovering them gives the number of models.
4535

46-
To see only labelings generated by our method : hamburger button of the 'method' column → select Equals 'Ours'
47-
([Set Filters](https://www.ag-grid.com/javascript-data-grid/filter-set/) are not available in the community edition of the library)
36+
## Column filters
4837

49-
To see only invalid labelings : hamburger button of the 'valid' column → choose 'False'
38+
To see only labelings generated by our method : hamburger button of the "method" column → select Equals "Ours"
39+
([Set Filters](https://www.ag-grid.com/javascript-data-grid/filter-set/) are not available in the community edition of the library).
5040

51-
To see only failed labeling generations : for ex. hamburger button of the '#charts' column → choose 'Blank'. To reset, choose 'Equals' and leave empty
41+
To see only invalid labelings : hamburger button of the "valid" column → choose "False".
5242

53-
# Column sort
43+
To see only failed labeling generations : e.g. hamburger button of the "#charts" column → choose "Blank". To reset, choose "Equals" and leave the field empty.
5444

55-
You can click on the name of a column to switch between ascending, descending and default order.
45+
## Column sort
5646

57-
# Columns description
47+
You can click on the name of a column to switch between ascending, descending and default order.
5848

59-
- CAD model : name of the 3D model
49+
## Columns description
6050

61-
- method : if the polycube labeling was generated using Evocube (Dumery et al. 2022) or our algorithm
51+
- CAD model : name of the 3D model. Bx, Sx and Mx models are from the MAMBO dataset, the others from OctreeMeshing
52+
- method : if the polycube labeling was generated using a graph-cut optimization (our initial labeling), _PolyCut_[^polycut], _Evocube_[^evocube] or our method
6253

6354
"surface mesh" group
6455

6556
- #vertices : number of vertices in the triangle mesh
66-
6757
- #facets : number of facets in the triangle mesh
68-
6958
- sd(area) : standard deviation of the facets area
7059

7160
"labeling" group
7261

7362
- #charts : number of charts
74-
7563
- #boundaries : number of boundaries
76-
7764
- #corners : number of corners
78-
7965
- #invalid-charts : number of invalid charts
80-
8166
- #invalid-boundaries : number of invalid boundaries
82-
8367
- #invalid-corners : number of invalid corners
84-
8568
- min(fidelity) : minimal (worse) per-triangle fidelity (closeness between the facet normal and the assigned direction). In [0,1]
86-
8769
- avg(fidelity) : average per-triangle fidelity, in [0,1]
88-
8970
- valid : if the labeling corresponds to a polycube topology or not
90-
9171
- #turning-points : number of turning-points (high turns along boundaries)
92-
93-
- duration : execution duration, of the labeling algorithm only, in seconds
94-
95-
- speedup : duration of Evocube divided by the duration of our algorithm
96-
97-
- open : open the 3D viewer. After 5s, the 3D model switches between the labeled mesh and a fast polycube estimation. If no labeling was generated, the triangle mesh is displayed instead. Screen space Ambient Occlusion (SSAO) is on by default.
72+
- duration : duration of the labeling stage with file I/O, in seconds. For our method, the generation of the initial labeling with graph-cut is
73+
- relative duration : duration divided by the duration of our method
74+
- open : open the 3D viewer. After 5s, the 3D model switches between the labeled mesh and a fast polycube estimation. If no labeling was generated, the triangle mesh is displayed instead. Screen Space Ambient Occlusion (SSAO) is on by default.
9875

9976
"feature edges" group
10077

101-
- removed : percentage of CAD lines not considered feature edges because of their low dihedral angle
102-
103-
- lost : percentage of lost feature edges because the same direction (label) is assigned on the 2 adjacent facets
104-
105-
- preserved : percentage of preserved feature edge (between 2 different labels)
78+
- ignored : percentage of CAD lines not considered feature edges because of their low dihedral angle
79+
- sharp & lost : percentage of lost feature edges because the same direction (label) is assigned on the 2 adjacent facets
80+
- sharp & preserved : percentage of preserved feature edge (between 2 different labels)
10681

10782
"hex-mesh" group
10883

10984
- min(SJ) : minimal (worse) per-hexahedron scaled jacobian
110-
11185
- avg(SJ) : average per-hexahedron scaled jacobian
86+
- open : open the 3D viewer. Screen Space Ambient Occlusion (SSAO) is off by default.
11287

113-
- open : open the 3D viewer. Screen space Ambient Occlusion (SSAO) is off by default.
114-
115-
# Comparison with PolyCut
116-
117-
We wanted to compare our algorithm with PolyCut (Livesu et al. 2013).
118-
However, we were unable to successfully run the executable provided by the authors (polycut.exe).
119-
We did not get any output file and the logs were empty.
120-
Following executables in pipeline.bat failed and printed "Input file missing".
121-
We contacted the corresponding author and did not get an answer before the submission deadline.
122-
123-
# Acknowledgment
88+
## Acknowledgment
12489

12590
The authors would like to thank contributors to the [AG Grid](https://github.com/ag-grid/ag-grid),
12691
[d3-sankey](https://github.com/d3/d3-sankey), [model-viewer](https://github.com/google/model-viewer)
12792
and [tinygltf](https://github.com/syoyo/tinygltf) libraries, as well as Khronos Group for the
12893
standardization of the [glTF](https://www.khronos.org/gltf/) file format.
12994

130-
Like our algorithm, data processing and the report generation will be made open-source.
95+
Like our algorithm, data processing and the report generation will be made open-source.
96+
97+
---
98+
99+
[^polycut]: M. Livesu, N. Vining, A. Sheffer, J. Gregson and R. Scateni, « PolyCut: Monotone Graph-Cuts for PolyCube Base-Complex Construction », in _Proceedings of SIGGRAPH Asia_, ACM, 2013. doi:[10.1145/2508363.2508388](https://doi.org/10.1145/2508363.2508388)
100+
101+
[^evocube]: C. Dumery, F. Protais, S. Mestrallet, C. Bourcier and F. Ledoux, « Evocube: a Genetic Labeling Framework for Polycube-Maps », vol. 41, nᵒ 6, p. 467‑479, 2022, doi: [10.1111/cgf.14649](https://doi.org/10.1111/cgf.14649)

0 commit comments

Comments
 (0)