Datasets:
TUB-synthetic
Synthetic turbid counterparts of the TUB (Turbid Underwater Baseline) dataset, released with the paper "Bridging the turbidity gap with synthetic data" (Ismiroglou, Pedersen, Bengtson, Moeslund; Pattern Analysis and Applications, 2026).
The paper extends the ICCV 2025 workshop paper Sea-ing Through Scattered Rays. The journal version evaluates the resulting data quantitatively and shows that training on STSR images improves generalization in real turbid scenes.
- π Paper: https://doi.org/10.1007/s10044-026-01770-4 (open access, CC BY 4.0)
- π Project page: https://vap.aau.dk/sea-ing-through-scattered-rays
- π Real turbid dataset: vapaau/TUB
Repository structure
baseline/ and STSR/ hold the main synthetic datasets. Each comes in two versions, one per monocular depth estimator used to produce z(x):
- Depth Anything V2 (
depthanythingv2/) - ZoeDepth (
zoedepth/)
baseline/ # reference IFM
βββ depthanythingv2/ # depth from Depth Anything V2
βββ zoedepth/ # depth from ZoeDepth
STSR/ # full STSR pipeline
βββ depthanythingv2/
βββ zoedepth/
ablations/
βββ baseline_standard_depth/ # baseline, fixed depth span between z_min and z_max
β βββ depthanythingv2/
β βββ zoedepth/
βββ STSR_standard_depth/ # STSR, fixed depth span between z_min and z_max
β βββ depthanythingv2/
β βββ zoedepth/
βββ only_inhomogeneity/ # baseline + GRF inhomogeneity only
β βββ depthanythingv2/
β βββ zoedepth/
βββ only_forwardscattering/ # baseline + forward scattering only
β βββ depthanythingv2/
β βββ zoedepth/
βββ baseline_no_depth/ # baseline, no depth map
βββ STSR_no_depth/ # STSR, no depth map
Per-variant layout
Each leaf folder holds 1,317 samples across 33 scenes (Exp1βExp33) and 4 cameras (Cam1βCam4). File stems match across the three subfolders and mirror TUB's Exp<scene>_Cam<view>_<idx> naming:
<variant>/
βββ images/ # Exp10_Cam1_001.png β synthetic turbid image, 1920Γ1080 RGB
βββ labels/ # Exp10_Cam1_001.txt β instance segmentation labels
βββ image_metadata/ # Exp10_Cam1_001.json β IFM parameters used for this image
Labels use the YOLO segmentation format, one object per line:
<class_id> x1 y1 x2 y2 ... xn yn # polygon vertices normalized to [0, 1]
Metadata stores the parameters used to render each image:
| Key | Description |
|---|---|
a |
absorption coefficient a(Ξ»), one value per sampled wavelength |
b |
scattering coefficient b(Ξ») |
beta |
beam attenuation coefficient Ξ²(Ξ») = a(Ξ») + b(Ξ») |
K |
diffuse downwelling attenuation coefficient K_d(Ξ») |
phi |
forward-scattering blur scale (Ο = Ο Β· z); null when forward scattering is disabled |
g |
fraction of scattered light leaving the line of sight, used in G = a + g Β· b |
mu |
backlight brightness factor (Bβ = ΞΌ Β· b Β· E / G) |
z_min, z_max |
metric scene-distance range (m) that the estimated depth map is rescaled to |
gamma |
gamma applied to the relative depth map before rescaling |
grf_min, grf_max |
range of the Gaussian Random Field that models medium inhomogeneity |
Citation
If you use this data, please cite:
@article{ismiroglou2026bridging,
title = {Bridging the turbidity gap with synthetic data},
author = {Ismiroglou, Vasiliki and Pedersen, Malte and Bengtson, Stefan H. and Moeslund, Thomas B.},
journal = {Pattern Analysis and Applications},
volume = {29},
number = {4},
pages = {186},
year = {2026},
publisher = {Springer},
doi = {10.1007/s10044-026-01770-4}
}
- Downloads last month
- 316