Skip to content

Commit 5c2e4ba

Browse files
chaofengcchaofengc
committed
docs: update model card
1 parent 5b5afb3 commit 5c2e4ba

File tree

1 file changed

+14
-25
lines changed

1 file changed

+14
-25
lines changed

‎docs/ModelCard.md

+14-25
Original file line numberDiff line numberDiff line change
@@ -13,7 +13,7 @@ print(pyiqa.list_models())
1313
| TOPIQ | `topiq_fr`, `topiq_fr-pipal` | Proposed in [this paper](https://arxiv.org/abs/2308.03060) |
1414
| AHIQ | `ahiq` |
1515
| PieAPP | `pieapp` |
16-
| LPIPS | `lpips`, `lpips-vgg`, `stlpips`, `stlpips-vgg` |
16+
| LPIPS | `lpips`, `lpips-vgg`, `stlpips`, `stlpips-vgg`, `lpips+`, `lpips-vgg+` |
1717
| DISTS | `dists` |
1818
| WaDIQaM | | *No pretrain models* |
1919
| CKDN<sup>[1](#fn1)</sup> | `ckdn` |
@@ -30,13 +30,15 @@ print(pyiqa.list_models())
3030

3131
| NR Method | Model names | Description |
3232
| ---------------------------- | ------------------------ |-------------------------------------------------------------------------------------|
33-
| ARNIQA | `arniqa`, `arniqa-live`, `arniqa-csiq`, `arniqa-tid`, `arniqa-kadid`, `arniqa-koniq`, `arniqa-clive`, `arniqa-flive`, `arniqa-spaq` | [ARNIQA](https://arxiv.org/abs/2310.14918) with different datasets, `koniq` by default |
33+
| Q-Align | `qalign` (with quality[default], aesthetic options) | Large vision-language models |
34+
| LIQE | `liqe`, `liqe_mix` | CLIP based method |
35+
| ARNIQA | `arniqa`, `arniqa-live`, `arniqa-csiq`, `arniqa-tid`, `arniqa-kadid`, `arniqa-clive`, `arniqa-flive`, `arniqa-spaq` | [ARNIQA](https://arxiv.org/abs/2310.14918) with different datasets, `koniq` by default |
3436
| TOPIQ | `topiq_nr`, `topiq_nr-flive`, `topiq_nr-spaq` | [TOPIQ](https://arxiv.org/abs/2308.03060) with different datasets, `koniq` by default |
35-
| TReS | `tres`, `tres-koniq`, `tres-flive` | TReS with different datasets, `koniq` by default |
37+
| TReS | `tres`, `tres-flive` | TReS with different datasets, `koniq` by default |
3638
| FID | `fid` | Statistic distance between two datasets |
3739
| CLIPIQA(+) | `clipiqa`, `clipiqa+`, `clipiqa+_vitL14_512`,`clipiqa+_rn50_512` | CLIPIQA(+) with different backbone, RN50 by default |
38-
| MANIQA | `maniqa`, `maniqa-kadid`, `maniqa-koniq`, `maniqa-pipal` | MUSIQ with different datasets, `koniq` by default |
39-
| MUSIQ | `musiq`, `musiq-koniq`, `musiq-spaq`, `musiq-paq2piq`, `musiq-ava` | MUSIQ with different datasets, `koniq` by default |
40+
| MANIQA | `maniqa`, `maniqa-kadid`, `maniqa-pipal` | MUSIQ with different datasets, `koniq` by default |
41+
| MUSIQ | `musiq`, `musiq-spaq`, `musiq-paq2piq`, `musiq-ava` | MUSIQ with different datasets, `koniq` by default |
4042
| DBCNN | `dbcnn` |
4143
| PaQ-2-PiQ | `paq2piq` |
4244
| HyperIQA | `hyperiqa` |
@@ -48,6 +50,7 @@ print(pyiqa.list_models())
4850
| BRISQUE | `brisque` | No backward |
4951
| ILNIQE | `ilniqe` | No backward |
5052
| NIQE | `niqe` | No backward |
53+
| PIQE | `piqe` | No backward |
5154
<!-- </tr>
5255
</table> -->
5356

@@ -61,25 +64,11 @@ print(pyiqa.list_models())
6164
| Face IQA | `topiq_nr-face` | TOPIQ model trained with face IQA dataset (GFIQA) |
6265
| Underwater IQA | `uranker` | A ranking-based underwater image quality assessment (UIQA) method, AAAI2023, [Arxiv](https://arxiv.org/abs/2208.06857), [Github](https://github.com/RQ-Wu/UnderwaterRanker) |
6366

64-
## Outputs of Different Metrics
67+
## Metric Output Score Range
6568
**Note: `~` means that the corresponding numeric bound is typical value and not mathematically guaranteed**
6669

67-
| model | lower better ? | min | max | DATE | Link |
68-
| -------- | -------------- | --- | ------- | ---- | --------------------------------------------------------------------------------------------------------------------------------------------------------- |
69-
| clipiqa | False | 0 | 1 | 2022 | https://arxiv.org/abs/2207.12396 |
70-
| maniqa | False | 0 | | 2022 | https://arxiv.org/abs/2204.08958 |
71-
| hyperiqa | False | 0 | 1 | 2020 | [pdf](https://openaccess.thecvf.com/content_CVPR_2020/papers/Su_Blindly_Assess_Image_Quality_in_the_Wild_Guided_by_a_CVPR_2020_paper.pdf) |
72-
| cnniqa | False | | | 2014 | [pdf](https://openaccess.thecvf.com/content_cvpr_2014/papers/Kang_Convolutional_Neural_Networks_2014_CVPR_paper.pdf) |
73-
| tres | False | | | 2022 | https://github.com/isalirezag/TReS |
74-
| musiq | False | ~0 | ~100 | 2021 | https://arxiv.org/abs/2108.05997 |
75-
| musiq-ava | False | ~0 | ~10 | 2021 | https://arxiv.org/abs/2108.05997 |
76-
| musiq-koniq | False | ~0 | ~100 | 2021 | https://arxiv.org/abs/2108.05997 |
77-
| musiq | False | | | 2021 | https://arxiv.org/abs/2108.05997 |
78-
| paq2piq | False | | | 2020 | [pdf](https://openaccess.thecvf.com/content_CVPR_2020/papers/Ying_From_Patches_to_Pictures_PaQ-2-PiQ_Mapping_the_Perceptual_Space_of_CVPR_2020_paper.pdf) |
79-
| dbcnn | False | | | 2019 | https://arxiv.org/bas/1907.02665 |
80-
| brisque | True | | | 2012 | [pdf](https://live.ece.utexas.edu/publications/2012/TIP%20BRISQUE.pdf) |
81-
| pi | True | | | 2018 | https://arxiv.org/abs/1809.07517 |
82-
| nima | False | | | 2018 | https://arxiv.org/abs/1709.05424 |
83-
| nrqm | False | | | 2016 | https://arxiv.org/abs/1612.05890 |
84-
| ilniqe | True | 0 | | 2015 | [pdf](http://www4.comp.polyu.edu.hk/~cslzhang/paper/IL-NIQE.pdf) |
85-
| niqe | True | 0 | | 2012 | [pdf](https://live.ece.utexas.edu/publications/2013/mittal2013.pdf) |
70+
You can now access the **rough** output range of each metric like this:
71+
```
72+
metric = pyiqa.create_metric('lpips')
73+
print(metric.score_range)
74+
```

0 commit comments

Comments
 (0)