Automated interpretation of side-scan sonar (SSS) imagery is traditionally hindered by view-dependence, as recorded intensities inextricably conflate intrinsic seabed reflectivity with transient acquisition geometries and acoustic artifacts. To address this, we introduce BenthicDINO, a physics-informed self-supervised learning framework that decouples acquisition dynamics from seabed representations. Utilizing a ConvNeXt-v2 backbone, physics-guided augmentations, and an HSIC penalty, BenthicDINO explicitly enforces feature invariance to physical viewing parameters like slant range and incidence angle. Evaluations demonstrate exceptional data efficiency - achieving near-peak performance in benthic habitat mapping with only 10% of available labels - and strong cross-domain generalization, outperforming fully supervised baselines on the AI4Shipwrecks benchmark using frozen representations.
Figures & tables
Fig. 1: SSS waterfall image showing a heading reversal midway through the survey. Blue and green boxes highlight the same physical seabed features imaged from opposite views. Vertical red dashed lines indicate 90∘ clockwise turns executed by the vehicle.
Fig. 2: ConvNeXt-v2 adapted for DINOv3. Outputs from stages 2–4 are upsampled to match stage 1, concatenated, and compressed via a fusion MLP into a 768-D embedding.
Fig. 3: Effects of TVG Attenuation on five sample SSS patches, progressing from the original patches (top row) to moderate (middle row) and extreme (bottom row) attenuation levels.
Fig. 4: The log of singular values against the singular value rank index of the fused representation.
Fig. 5: Distribution of feature invariance across augmented pairs. The dashed line indicates a mean cosine similarity of 0.914 , with half of the data points falling below this line and the other half densely concentrated above it.
Fig. 6: PCA Visualizations of feature maps across network stages extracted from Synthetic Aperature Sonar (SAS) images from the SASSED dataset [ 45 ] .
Evaluator
Metric
1% Data
5% Data
10% Data
50% Data
100% Data
Linear Probe
Accuracy
0.860
0.919
0.930
0.946
0.949
Macro F1
0.825
0.897
0.914
0.932
0.936
mIoU
0.706
0.817
0.844
0.876
0.882
K-NN
Accuracy
0.841
0.886
0.904
0.927
0.935
Macro F1
0.762
0.836
0.865
0.899
0.912
mIoU
0.632
0.727
0.769
0.821
0.842
TABLE I: Few-Shot Generalization: Linear Probe vs. K-NN using the Fused Representation
Representation
Accuracy
Macro F1
mIoU
Stage 1
0.686
0.578
0.434
Stage 2
0.815
0.757
0.627
Stage 3
0.938
0.920
0.855
Stage 4
0.932
0.917
0.849
Fused Representation
0.949
0.936
0.882
TABLE II: Stage-by-Stage Feature Progression (100% Data, Linear Probe)
Benthic Class
Precision
Recall
F1-Score
IoU
Sand Ripples
0.960
0.957
0.959
0.920
Fine Sediments
0.984
0.974
0.979
0.959
Rocks
0.847
0.893
0.869
0.769
Maerl
0.948
0.925
0.937
0.881
TABLE III: Per-Class Performance for the Fused Representation (100% Data, Linear Probe)
Method
IoU ship
F1
TPR
TNR
Yang et al. [ 49 ]
0.212
0.310
0.296
0.995
ViT-Adapter [ 50 ]
0.283
0.395
0.488
0.995
DeepLabv3 [ 51 ]
0.363
0.473
0.485
0.996
HRNet [ 52 ]
0.372
0.490
0.516
0.998
U-Net [ 53 ]
0.411
0.526
0.592
0.997
SOD (InSPyReNet) [ 54 ]
0.445
0.594
0.652
0.997
TABLE IV: Site-averaged shipwreck segmentation performance on the AI4Shipwrecks test split.
Site
Depth
Ours
Best baseline
Best
(m)
(model)
W.H. Gilbert
77.7
0.655
SOD
0.749
Viator
57.3
0.729
SOD
0.776
Corsair
55.5
0.694
SOD
0.583
Corsican
48.8
0.566
HRNet
0.174
Monrovia
42.7
0.612
U-Net
0.572
TABLE V: Per-site IoU ship on the 13 wreck-bearing test sites, sorted by depth in descending order.
Fig. 7: Mean IoU ship on the AI4Shipwrecks test split, stratified by site depth.
Fig. 8: Segmentation results from the BenthicDINO encoder + U-Net head applied to the AI4Shipwrecks dataset [ 55 ] . Left to right, original SSS image, ground truth annotation, BenthicDINO result.
Fig. 9: PCA visualization of feature maps across network stages. The first three principal components are mapped to RGB channels to visualize the high-dimensional feature spaces. Rows (from top to bottom) display: original image patches, intermediate outputs from Stages 1 through 4, and the final fused representation.
Fig. 10: Top five matches for a sample of the centroids obtained by running K-Means clustering on 56 clusters. The centroid label is outlined on top of each respective image.
Fig. 11: Continued: Top five matches for a sample of the centroids obtained using the K-Means clustering.