nihermann nielsr HF Staff commited on
Commit
6a90b42
·
verified ·
1 Parent(s): b6539da

Improve dataset card: Add metadata, links, and tags for discoverability (#2)

Browse files

- Improve dataset card: Add metadata, links, and tags for discoverability (f39a3dbac1dca492e0b3ea9995f842022c3f0f91)
- Changed some formatting and removed additional links to arXiv and the project page (700b4b92e166ff6e124db18db29cbd480ea14eb7)


Co-authored-by: Niels Rogge <[email protected]>

Files changed (1) hide show
  1. README.md +24 -17
README.md CHANGED
@@ -1,20 +1,36 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
  # <img src="https://www.svgrepo.com/show/510149/puzzle-piece.svg" width="22"/> Puzzle Similarity
2
 
3
- -----
4
 
5
- > This repository contains the dataset presented in the ICCV 2025 paper "Puzzle Similarity: A Perceptually-guided Cross-Reference Metric for Artifact Detection in 3D Scene Reconstructions" (See arxiv.org/abs/2411.17489 or the [project page](https://nihermann.github.io/))
6
- > Authors: Nicolai Hermann, Jorge Condor, and Piotr Didyk
7
 
 
 
8
 
9
  ### Dataset Description
10
  The Dataset consists of 36 hand-selected 3D Gaussian Splatting renderings containing common reconstruction artefacts, ground truths, human-annotated masks, and a set of reference views of the same scene.
11
 
12
  Each mask is an average of 22 binary masks, each created by a different human participant who was asked to annotate areas in the reconstructed images that they perceived as visually degraded, unnatural, or incongruent. The dataset can be used to benchmark No-Reference, Cross-Reference, and Full-Reference image quality metrics for their correlation with human judgment. The naming convention of the data is as follows:
13
 
14
- dataset_perc_id_mask.png (grayscale)
15
- dataset_perc_id_artifact.png
16
- dataset_perc_id_gt.png
17
- dataset_perc_refs/
18
 
19
  The dataset was created by fitting 3DGS to a scene while using a reduced number of training views. We withheld a percentage of views (perc) and added them to the validation dataset, which is found in the *_refs/ directory for each respective sample to act as unseen reference views for Cross-Reference metrics. We fitted the scenes while withholding 60%, 70%, or 80% to get a wider variety and strength of artifacts. (Disclaimer: perc actually refers to proportions, so the possible values are 0.6, 0.7, or 0.8)
20
 
@@ -26,7 +42,6 @@ The included datasets are a collection from the Mip-NeRF360 [1], Tanks and Templ
26
 
27
  [3] Peter Hedman, Julien Philip, True Price, Jan-Michael Frahm, 619 George Drettakis, and Gabriel Brostow. Deep blending for 620 free-viewpoint image-based rendering. ACM Transactions 621 on Graphics, 37(6):1–15, 2018.
28
 
29
-
30
  ### Citation
31
  If you find this work useful, please consider citing:
32
  ```bibtex
@@ -39,12 +54,4 @@ If you find this work useful, please consider citing:
39
  primaryClass={cs.CV},
40
  url={https://arxiv.org/abs/2411.17489},
41
  }
42
- ```
43
-
44
- ---
45
- license: apache-2.0
46
- language:
47
- - en
48
- size_categories:
49
- - 1K<n<10K
50
- ---
 
1
+ ---
2
+ license: apache-2.0
3
+ language:
4
+ - en
5
+ size_categories:
6
+ - 1K<n<10K
7
+ task_categories:
8
+ - image-segmentation
9
+ tags:
10
+ - 3d-reconstruction
11
+ - artifact-detection
12
+ - image-quality-assessment
13
+ - human-annotation
14
+ ---
15
+
16
  # <img src="https://www.svgrepo.com/show/510149/puzzle-piece.svg" width="22"/> Puzzle Similarity
17
 
18
+ [Project page](https://nihermann.github.io/puzzlesim/) | [Paper](https://arxiv.org/abs/2411.17489) | [Code](https://github.com/nihermann/PuzzleSim)
19
 
20
+ -----
 
21
 
22
+ > This repository contains the dataset presented in the ICCV 2025 paper "Puzzle Similarity: A Perceptually-guided Cross-Reference Metric for Artifact Detection in 3D Scene Reconstructions"
23
+ > Authors: Nicolai Hermann, Jorge Condor, and Piotr Didyk
24
 
25
  ### Dataset Description
26
  The Dataset consists of 36 hand-selected 3D Gaussian Splatting renderings containing common reconstruction artefacts, ground truths, human-annotated masks, and a set of reference views of the same scene.
27
 
28
  Each mask is an average of 22 binary masks, each created by a different human participant who was asked to annotate areas in the reconstructed images that they perceived as visually degraded, unnatural, or incongruent. The dataset can be used to benchmark No-Reference, Cross-Reference, and Full-Reference image quality metrics for their correlation with human judgment. The naming convention of the data is as follows:
29
 
30
+ - `dataset_perc_id_mask.png` (grayscale)
31
+ - `dataset_perc_id_artifact.png`
32
+ - `dataset_perc_id_gt.png`
33
+ - `dataset_perc_refs/`
34
 
35
  The dataset was created by fitting 3DGS to a scene while using a reduced number of training views. We withheld a percentage of views (perc) and added them to the validation dataset, which is found in the *_refs/ directory for each respective sample to act as unseen reference views for Cross-Reference metrics. We fitted the scenes while withholding 60%, 70%, or 80% to get a wider variety and strength of artifacts. (Disclaimer: perc actually refers to proportions, so the possible values are 0.6, 0.7, or 0.8)
36
 
 
42
 
43
  [3] Peter Hedman, Julien Philip, True Price, Jan-Michael Frahm, 619 George Drettakis, and Gabriel Brostow. Deep blending for 620 free-viewpoint image-based rendering. ACM Transactions 621 on Graphics, 37(6):1–15, 2018.
44
 
 
45
  ### Citation
46
  If you find this work useful, please consider citing:
47
  ```bibtex
 
54
  primaryClass={cs.CV},
55
  url={https://arxiv.org/abs/2411.17489},
56
  }
57
+ ```