Datasets:

3L-OPTRONICS
/

3L-Dataset

---
pretty_name: 3L-Dataset · Low Light Level
license: other
license_name: 3l-dataset-research-license
license_link: LICENSE.md
extra_gated_heading: "Access the 3L-Dataset sample"
extra_gated_description: "The 3L-Dataset is released for non-commercial research only. Please read the license and fill in the form below. Access is granted after review."
extra_gated_button_content: "I accept the license and request access"
extra_gated_prompt: "By requesting access, you agree to the 3L-Dataset Research License: non-commercial research use only, no redistribution of the images (original or modified), and citation of the paper in any publication."
extra_gated_fields:
  Full name: text
  Affiliation (university, laboratory or company): text
  Professional email: text
  Country: country
  Intended use: text
  I will use the dataset for non-commercial research only: checkbox
  I will not redistribute the images, in original or modified form: checkbox
  I will cite the 3L-Dataset paper in any resulting publication: checkbox
language:
  - en
task_categories:
  - image-to-image
  - image-classification
tags:
  - low-light
  - low-light-image-enhancement
  - llie
  - night-vision
  - image-restoration
  - paired-images
  - benchmark
size_categories:
  - n<1K                  # TODO: adjust to the real number of images (1K<n<10K, ...)
---
# 3L-Dataset · Low Light Level

**An aligned, physically realistic, low-light dataset.**

Developed by [3L-OPTRONICS](https://3l-optronics.com), the 3L-Dataset provides paired captures of the same scenes under standardized nighttime illumination levels, for training and evaluating **Low-Light Image Enhancement (LLIE)** methods.

| | |
|---|---|
| 🌐 **Website** | https://3l-d.github.io/3L-Dataset/ |
| 📄 **Paper** | *3L-D: A Low-Light Level Dataset from a Sensor-Agnostic, Alignment-Preserving Acquisition Framework* (pending, arXiv link coming soon) |
| ✉️ **Contact** | contact@3l-optronics.com |
| 🤗 **Dataset** | https://huggingface.co/datasets/3L-OPTRONICS/3L-Dataset |
| 💼 **LinkedIn** | https://www.linkedin.com/company/3l-optronics |
| ⚖️ **License** | [3L-Dataset Research License](LICENSE.md): non-commercial research only |

> This repository hosts a **sample** of the dataset. See the paper, *3L-D: A Low-Light Level Dataset from a Sensor-Agnostic, Alignment-Preserving Acquisition Framework*, for the full acquisition protocol.

### ⚠️ License at a glance

This dataset is released under the **[3L-Dataset Research License](LICENSE.md)**. By downloading it, you agree to its terms:

- ✅ **Non-commercial research and teaching only**
- ❌ **No commercial use** of the dataset or of models trained on it
- ❌ **No redistribution** of the images, in original or modified form
- 📝 **Cite the paper** in any publication using the dataset

Full terms in [License](#license) and [LICENSE.md](LICENSE.md). Commercial use: **contact@3l-optronics.com**

---

## Why this dataset

Most low-light datasets rely on synthetic darkening or on handheld captures that are never perfectly aligned. The - is captured on a **calibrated optical bench**:

- **Physically realistic**: darkness is produced by real, measured illumination, not simulated in post-processing.
- **Pixel-perfect alignment**: camera and scene stay fixed between levels, so every low-light image has an exactly aligned reference.
- **Reproducible**: each acquisition follows the same controlled lighting conditions.

Because full-reference metrics such as **PSNR, SSIM and LPIPS** compare images pixel by pixel, this alignment makes their scores interpretable: they measure the quality of the restoration, not a camera shift.

## Acquisition setup

- Light-tight enclosure with an **integrating sphere** for uniform, calibrated illumination
- **Rotating stage** to capture each scene from several angles
- Scale-model scenes reproducing outdoor environments (countryside, forest, desert, urban)
- Fixed camera and optics (f/1.2 objective)
- **6 illumination levels** per scene and per position: one reference level (`gt`) and five night levels (`n1` to `n5`)

## Dataset structure

This sample contains the reference level and three night levels:

```
3l-dataset/
├── metadata.jsonl      ← one line per image, read by load_dataset
├── gt/                 ← reference (ground truth)
│   ├── metadata.json
│   ├── 0001.png
│   └── ...
├── n2/
│   ├── metadata.json
│   └── ...
├── n3/
└── n4/
```

| Folder | Content |
|---|---|
| `gt` | **Ground truth**: reference capture of each scene and position, used as the target for full-reference metrics |
| `n2`, `n3`, `n4` | The same scenes and positions under increasingly low night illumination |

`gt` is the reference level. Night levels `n1` and `n5` are part of the full dataset but not of this sample. For a given scene and position, the captures of all levels are aligned pixel by pixel.


| Block | Content | Purpose |
|---|---|---|
| **Images** | Low-light paired acquisitions | Provide the raw material |
| **Classes** | One folder per illumination level (`gt`, `n2`, `n3`, `n4`): the folder name is the label | Enable supervised training and evaluation |
| **Metadata** | One JSON file per folder: image name and stage angle (degrees) for each image, plus session time, integration time, start and end angles. The same information is gathered in `metadata.jsonl` at the root | Make results comparable and reproducible |

**Image format:** 16-bit grayscale PNG, 2808 × 1096 px.

16-bit depth preserves the full dynamic range of the sensor, which matters in very low light where most of the signal sits in the lowest grey levels. Load the images without converting them to 8-bit:

```python
import numpy as np
from PIL import Image
img = cv2.imread("gt/0001.png",-1)  # dtype uint16, shape (1096, 2808)
img = img.astype(np.float32) / 65535.0       # normalized to [0, 1]
```

### Metadata example

Excerpt of `gt/metadata.json`:


```json
{
  "session_time": "2026-08-21T10:00",
  "integration_time_ms": 120,
  "start_angle_deg": 0,
  "end_angle_deg": 360,
  "images": [
    { "image": "gt/0001.png", "angle_deg": 0 },
    { "image": "gt/0011.png", "angle_deg": 150 }
  ]
}
```

## Usage

```python
from datasets import load_dataset
ds = load_dataset("3L-OPTRONICS/3L-Dataset")
print(ds)
```

This dataset is **gated**: request access on this page first, then log in once from your terminal before loading it:

```bash
hf auth login
```

`load_dataset` reads `metadata.jsonl` automatically: every image comes with its illumination level (`level`), its stage angle and its session information.

```python
sample = ds["train"][0]
sample["image"]        # PIL image
sample["level"]        # "gt", "n2", "n3" or "n4"
sample["angle_deg"]    # stage angle in degrees
```


## Benchmark

Eight LLIE networks evaluated on the 3L-Dataset (see the paper for the full protocol).

| Network | PSNR ↑ | SSIM ↑ | LPIPS ↓ | NIQE ↓ | Inference time (ms) ↓ |
|---|---:|---:|---:|---:|---:|
| SCI | 37.68 | 0.8673 | 0.6905 | 29.07 | **2.08** |
| EnlightenGAN | 36.66 | 0.8676 | 0.0801 | 24.04 | 58.14 |
| Zero-DCE | 37.90 | 0.8693 | 0.7785 | **10.47** | 26.21 |
| KinD | 41.98 | 0.9298 | 0.0277 | 16.01 | 87.22 |
| URetinex-Net++ | 40.95 | 0.9158 | 0.0311 | 20.87 | 254.01 |
| WaveMamba | **54.80** | **0.9808** | 0.0013 | 21.71 | 1322.63 |
| MambaIR | 51.85 | 0.9614 | **0.0011** | 21.36 | 9199.43 |
| RetinexFormer | 44.64 | 0.9687 | 0.0210 | 17.78 | 541.14 |

PSNR and SSIM are higher-is-better; LPIPS, NIQE and inference time are lower-is-better.

## Custom datasets and sensor characterization

The same calibrated bench can be reconfigured for your own requirements:

- **Custom datasets**: specific wavelengths (λ), visible low light, custom night levels, your optics and sensor.
- **Sensor characterization**: evaluate how well a sensor sees at night, with metrics computed on our standard night levels or on your own specifications.

Contact: **contact@3l-optronics.com**

## License

© 2026 3L-OPTRONICS. All rights reserved, except as expressly granted below.

The 3L-Dataset is released under the **[3L-Dataset Research License](LICENSE.md)**.

| | |
|---|---|
| ✅ **Allowed** | Download and use the images for **non-commercial research and teaching** |
| ✅ **Allowed** | Train and evaluate models for non-commercial research |
| ✅ **Allowed** | Show a **limited number** of images and processed results (e.g. enhanced outputs) **in scientific publications**, with citation |
| ❌ **Prohibited** | Any **commercial use**: products, services, paid consulting, training or benchmarking commercial models |
| ❌ **Prohibited** | **Redistributing** the dataset or any part of it, in original or **modified** form (crops, edits, re-encoding, mirrors, other dataset hubs) |
| ❌ **Prohibited** | Creating and sharing a **derived dataset** from these images |
| ❌ **Prohibited** | Removing copyright notices or misrepresenting the origin of the images |
| 📝 **Required** | **Citing** the 3L-Dataset paper in any publication or presentation using the dataset |

Access to this repository is **gated**: you must accept the license and provide your affiliation before downloading.

**Commercial use, redistribution or custom datasets** are available under a separate agreement: **contact@3l-optronics.com**

## Citation

If you use the 3L-Dataset, please cite:

```bibtex
@misc{rueff2026_3ld,
  title  = {{3L-D}: A Low-Light Level Dataset from a Sensor-Agnostic, Alignment-Preserving Acquisition Framework},
  author = {Rueff, L{\'e}o and Ghezal, Maurg{\`e}ne and Gorse, Valentin and Dufaure, Astrid and
            Monsalvo, Esteban and Atteia, Fran{\c{c}}ois and Escoubas, Ludovic},
  year   = {2026},
  note   = {Preprint, arXiv submission in preparation}
}
```

<!-- TODO: once on arXiv, add: eprint = {XXXX.XXXXX}, archivePrefix = {arXiv}, primaryClass = {cs.CV}, url = {https://arxiv.org/abs/XXXX.XXXXX} -->

---

© 2026 [3L-OPTRONICS](https://3l-optronics.com) · [Website](https://maurgene-ghezal-3l.github.io/3L-Dataset/) · Paper (pending) · [LinkedIn](https://www.linkedin.com/company/3l-optronics)

	---
	pretty_name: 3L-Dataset · Low Light Level
	license: other
	license_name: 3l-dataset-research-license
	license_link: LICENSE.md
	extra_gated_heading: "Access the 3L-Dataset sample"
	extra_gated_description: "The 3L-Dataset is released for non-commercial research only. Please read the license and fill in the form below. Access is granted after review."
	extra_gated_button_content: "I accept the license and request access"
	extra_gated_prompt: "By requesting access, you agree to the 3L-Dataset Research License: non-commercial research use only, no redistribution of the images (original or modified), and citation of the paper in any publication."
	extra_gated_fields:
	Full name: text
	Affiliation (university, laboratory or company): text
	Professional email: text
	Country: country
	Intended use: text
	I will use the dataset for non-commercial research only: checkbox
	I will not redistribute the images, in original or modified form: checkbox
	I will cite the 3L-Dataset paper in any resulting publication: checkbox
	language:
	- en
	task_categories:
	- image-to-image
	- image-classification
	tags:
	- low-light
	- low-light-image-enhancement
	- llie
	- night-vision
	- image-restoration
	- paired-images
	- benchmark
	size_categories:
	- n<1K # TODO: adjust to the real number of images (1K<n<10K, ...)
	---

	# 3L-Dataset · Low Light Level

	An aligned, physically realistic, low-light dataset.

	Developed by [3L-OPTRONICS](https://3l-optronics.com), the 3L-Dataset provides paired captures of the same scenes under standardized nighttime illumination levels, for training and evaluating Low-Light Image Enhancement (LLIE) methods.

	\| \| \|
	\|---\|---\|
	\| 🌐 Website \| https://3l-d.github.io/3L-Dataset/ \|
	\| 📄 Paper \| 3L-D: A Low-Light Level Dataset from a Sensor-Agnostic, Alignment-Preserving Acquisition Framework (pending, arXiv link coming soon) \|
	\| ✉️ Contact \| contact@3l-optronics.com \|
	\| 🤗 Dataset \| https://huggingface.co/datasets/3L-OPTRONICS/3L-Dataset \|
	\| 💼 LinkedIn \| https://www.linkedin.com/company/3l-optronics \|
	\| ⚖️ License \| [3L-Dataset Research License](LICENSE.md): non-commercial research only \|

	> This repository hosts a sample of the dataset. See the paper, 3L-D: A Low-Light Level Dataset from a Sensor-Agnostic, Alignment-Preserving Acquisition Framework, for the full acquisition protocol.

	### ⚠️ License at a glance

	This dataset is released under the [3L-Dataset Research License](LICENSE.md). By downloading it, you agree to its terms:

	- ✅ Non-commercial research and teaching only
	- ❌ No commercial use of the dataset or of models trained on it
	- ❌ No redistribution of the images, in original or modified form
	- 📝 Cite the paper in any publication using the dataset

	Full terms in [License](#license) and [LICENSE.md](LICENSE.md). Commercial use: contact@3l-optronics.com

	---

	## Why this dataset

	Most low-light datasets rely on synthetic darkening or on handheld captures that are never perfectly aligned. The - is captured on a calibrated optical bench:

	- Physically realistic: darkness is produced by real, measured illumination, not simulated in post-processing.
	- Pixel-perfect alignment: camera and scene stay fixed between levels, so every low-light image has an exactly aligned reference.
	- Reproducible: each acquisition follows the same controlled lighting conditions.

	Because full-reference metrics such as PSNR, SSIM and LPIPS compare images pixel by pixel, this alignment makes their scores interpretable: they measure the quality of the restoration, not a camera shift.

	## Acquisition setup

	- Light-tight enclosure with an integrating sphere for uniform, calibrated illumination
	- Rotating stage to capture each scene from several angles
	- Scale-model scenes reproducing outdoor environments (countryside, forest, desert, urban)
	- Fixed camera and optics (f/1.2 objective)
	- 6 illumination levels per scene and per position: one reference level (`gt`) and five night levels (`n1` to `n5`)

	## Dataset structure

	This sample contains the reference level and three night levels:

	```
	3l-dataset/
	├── metadata.jsonl ← one line per image, read by load_dataset
	├── gt/ ← reference (ground truth)
	│ ├── metadata.json
	│ ├── 0001.png
	│ └── ...
	├── n2/
	│ ├── metadata.json
	│ └── ...
	├── n3/
	└── n4/
	```

	\| Folder \| Content \|
	\|---\|---\|
	\| `gt` \| Ground truth: reference capture of each scene and position, used as the target for full-reference metrics \|
	\| `n2`, `n3`, `n4` \| The same scenes and positions under increasingly low night illumination \|

	`gt` is the reference level. Night levels `n1` and `n5` are part of the full dataset but not of this sample. For a given scene and position, the captures of all levels are aligned pixel by pixel.


	\| Block \| Content \| Purpose \|
	\|---\|---\|---\|
	\| Images \| Low-light paired acquisitions \| Provide the raw material \|
	\| Classes \| One folder per illumination level (`gt`, `n2`, `n3`, `n4`): the folder name is the label \| Enable supervised training and evaluation \|
	\| Metadata \| One JSON file per folder: image name and stage angle (degrees) for each image, plus session time, integration time, start and end angles. The same information is gathered in `metadata.jsonl` at the root \| Make results comparable and reproducible \|

	Image format: 16-bit grayscale PNG, 2808 × 1096 px.

	16-bit depth preserves the full dynamic range of the sensor, which matters in very low light where most of the signal sits in the lowest grey levels. Load the images without converting them to 8-bit:

	```python
	import numpy as np
	from PIL import Image

	img = cv2.imread("gt/0001.png",-1) # dtype uint16, shape (1096, 2808)
	img = img.astype(np.float32) / 65535.0 # normalized to [0, 1]
	```

	### Metadata example

	Excerpt of `gt/metadata.json`:


	```json
	{
	"session_time": "2026-08-21T10:00",
	"integration_time_ms": 120,
	"start_angle_deg": 0,
	"end_angle_deg": 360,
	"images": [
	{ "image": "gt/0001.png", "angle_deg": 0 },
	{ "image": "gt/0011.png", "angle_deg": 150 }
	]
	}
	```

	## Usage

	```python
	from datasets import load_dataset

	ds = load_dataset("3L-OPTRONICS/3L-Dataset")
	print(ds)
	```

	This dataset is gated: request access on this page first, then log in once from your terminal before loading it:

	```bash
	hf auth login
	```

	`load_dataset` reads `metadata.jsonl` automatically: every image comes with its illumination level (`level`), its stage angle and its session information.

	```python
	sample = ds["train"][0]
	sample["image"] # PIL image
	sample["level"] # "gt", "n2", "n3" or "n4"
	sample["angle_deg"] # stage angle in degrees
	```


	## Benchmark

	Eight LLIE networks evaluated on the 3L-Dataset (see the paper for the full protocol).

	\| Network \| PSNR ↑ \| SSIM ↑ \| LPIPS ↓ \| NIQE ↓ \| Inference time (ms) ↓ \|
	\|---\|---:\|---:\|---:\|---:\|---:\|
	\| SCI \| 37.68 \| 0.8673 \| 0.6905 \| 29.07 \| 2.08 \|
	\| EnlightenGAN \| 36.66 \| 0.8676 \| 0.0801 \| 24.04 \| 58.14 \|
	\| Zero-DCE \| 37.90 \| 0.8693 \| 0.7785 \| 10.47 \| 26.21 \|
	\| KinD \| 41.98 \| 0.9298 \| 0.0277 \| 16.01 \| 87.22 \|
	\| URetinex-Net++ \| 40.95 \| 0.9158 \| 0.0311 \| 20.87 \| 254.01 \|
	\| WaveMamba \| 54.80 \| 0.9808 \| 0.0013 \| 21.71 \| 1322.63 \|
	\| MambaIR \| 51.85 \| 0.9614 \| 0.0011 \| 21.36 \| 9199.43 \|
	\| RetinexFormer \| 44.64 \| 0.9687 \| 0.0210 \| 17.78 \| 541.14 \|

	PSNR and SSIM are higher-is-better; LPIPS, NIQE and inference time are lower-is-better.

	## Custom datasets and sensor characterization

	The same calibrated bench can be reconfigured for your own requirements:

	- Custom datasets: specific wavelengths (λ), visible low light, custom night levels, your optics and sensor.
	- Sensor characterization: evaluate how well a sensor sees at night, with metrics computed on our standard night levels or on your own specifications.

	Contact: contact@3l-optronics.com

	## License

	© 2026 3L-OPTRONICS. All rights reserved, except as expressly granted below.

	The 3L-Dataset is released under the [3L-Dataset Research License](LICENSE.md).

	\| \| \|
	\|---\|---\|
	\| ✅ Allowed \| Download and use the images for non-commercial research and teaching \|
	\| ✅ Allowed \| Train and evaluate models for non-commercial research \|
	\| ✅ Allowed \| Show a limited number of images and processed results (e.g. enhanced outputs) in scientific publications, with citation \|
	\| ❌ Prohibited \| Any commercial use: products, services, paid consulting, training or benchmarking commercial models \|
	\| ❌ Prohibited \| Redistributing the dataset or any part of it, in original or modified form (crops, edits, re-encoding, mirrors, other dataset hubs) \|
	\| ❌ Prohibited \| Creating and sharing a derived dataset from these images \|
	\| ❌ Prohibited \| Removing copyright notices or misrepresenting the origin of the images \|
	\| 📝 Required \| Citing the 3L-Dataset paper in any publication or presentation using the dataset \|

	Access to this repository is gated: you must accept the license and provide your affiliation before downloading.

	Commercial use, redistribution or custom datasets are available under a separate agreement: contact@3l-optronics.com

	## Citation

	If you use the 3L-Dataset, please cite:

	```bibtex
	@misc{rueff2026_3ld,
	title = {{3L-D}: A Low-Light Level Dataset from a Sensor-Agnostic, Alignment-Preserving Acquisition Framework},
	author = {Rueff, L{\'e}o and Ghezal, Maurg{\`e}ne and Gorse, Valentin and Dufaure, Astrid and
	Monsalvo, Esteban and Atteia, Fran{\c{c}}ois and Escoubas, Ludovic},
	year = {2026},
	note = {Preprint, arXiv submission in preparation}
	}
	```

	<!-- TODO: once on arXiv, add: eprint = {XXXX.XXXXX}, archivePrefix = {arXiv}, primaryClass = {cs.CV}, url = {https://arxiv.org/abs/XXXX.XXXXX} -->

	---

	© 2026 [3L-OPTRONICS](https://3l-optronics.com) · [Website](https://maurgene-ghezal-3l.github.io/3L-Dataset/) · Paper (pending) · [LinkedIn](https://www.linkedin.com/company/3l-optronics)