PaperlayerView exact revision on arXiv

Source-linked AI summary

EddyNet: A Deep Neural Network For Pixel-Wise Classification of Oceanic Eddies

Redouane Lguensat, Miao Sun, Ronan Fablet, Evan Mason, Pierre Tandeo, Ge Chen

arXiv:1711.03954v1cs.CVphysics.ao-ph

TL;DR

Automated eddy detection and classification from SSH maps requires a pixel-wise approach suited to ocean remote sensing. The paper proposes EddyNet, a U-Net-like segmentation network using activation and overlap-loss alternatives, and evaluates its classification performance, including ghost eddies.

  • Problem

    The paper addresses automated pixel-wise detection and classification of oceanic eddies from SSH maps.

    Source: §I · p. 1

  • Method

    EddyNet uses a U-Net-like deep segmentation architecture with SELU or ReLU plus Batch Normalization and Dice-based overlap loss alternatives.

    Source: §I · p. 1, B. Loss metric · p. 3

  • Results

    EddyNet assigns ghost-eddy centers to the correct classes 55% of the time for anticyclonic eddies and 45% for cyclonic eddies.

    Source: B. Ghost eddies · p. 4

  • Takeaways & Limitations

    The approach transfers deep-learning image-segmentation methods to ocean eddy classification and provides an alternative to conventional detection methods for ghost eddies.

    Source: §VI · p. 4, B. Ghost eddies · p. 4

Abstract

from arXiv · show

This work presents EddyNet, a deep learning based architecture for automated eddy detection and classification from Sea Surface Height (SSH) maps provided by the Copernicus Marine and Environment Monitoring Service (CMEMS). EddyNet is a U-Net like network that consists of a convolutional encoder-decoder followed by a pixel-wise classification layer. The output is a map with the same size of the input where pixels have the following labels \{'0': Non eddy, '1': anticyclonic eddy, '2': cyclonic eddy\}. We investigate the use of SELU activation function instead of the classical ReLU+BN and we use an overlap based loss function instead of the cross entropy loss. Keras Python code, the training datasets and EddyNet weights files are open-source and freely available on https://github.com/redouanelg/EddyNet.

I. INTRODUCTION

The paper presents EddyNet, a deep neural network for automated eddy detection and classification from SSH maps. It investigates SELU activations, Dice-based loss, and dropout within this image-segmentation setting.

  • EddyNet applies deep-learning image-segmentation methods to pixel-wise eddy detection and classification from AVISO-SSH maps.Source: §I · p. 1
  • The architecture investigates SELU activations as an alternative to ReLU with Batch Normalization.Source: §I · p. 1
  • Dice-based loss produces better scores for cyclonic and anticyclonic classes than categorical cross-entropy.Source: §I · p. 1
  • Dropout layers are added to prevent EddyNet from overfitting.Source: §I · p. 1
  • The authors describe the work as the first deep-learning architecture for pixel-wise eddy classification.Source: §I · p. 1

II. PROBLEM STATEMENT AND RELATED WORK

Oceanic eddies transport important physical quantities and influence ocean circulation, motivating automated detection. Existing approaches include physical-parameter, geometric-contour, hybrid, and machine-learning methods, while this work trains on PET14 outputs.

  • Oceanic eddies transport carbon, heat, phytoplankton, and salt, making their detection relevant to circulation, weather, mixing, and climate models.Source: §II · p. 1
  • Physical-parameter methods such as Okubo-Weiss are criticized for expert dependence, regional specificity, and sensitivity to noisy SSH maps.Source: §II · pp. 1–2
  • Geometric methods model eddies as elliptical shapes using closed contours, while hybrid methods combine physical and geometric ideas.Source: §II · p. 2
  • EddyNet uses PET14 eddy detection and classification outputs as its training database.Source: §II · p. 2

III. DATA PREPARATION

The dataset uses 15 years of daily PET14 detections paired with CMEMS SSH maps, sampled into 128 × 128 Southern Atlantic patches for training and testing.

  • The dataset contains 15 years of daily detected and classified eddies from 1998–2012, paired with AVISO-SSH maps from CMEMS.Source: §III · p. 2
  • Training uses the first 14 years, while 2012 is reserved for testing.Source: §III · p. 2
  • Memory constraints set the architecture input size to 128 × 128 pixels.Source: §III · p. 2
  • Randomly sampling one patch per SSH map produces 5100 training samples from the Southern Atlantic region.Source: §III · p. 2
  • Land and missing-data fill values are replaced by zero to avoid outliers without affecting eddy detection.Source: §III · p. 2

A. EddyNet architecture

EddyNet is a three-stage U-Net-like encoder-decoder with skip connections, contrasting SELU activation with ReLU plus Batch Normalization and using transposed convolutions for upsampling.

B. Loss metric

EddyNet uses a differentiable Dice-based overlap loss for multiclass segmentation, optimizing the mean of one-vs-all soft Dice coefficients across the three classes.

B. Loss metric

  • The Dice coefficient measures twice the predicted–ground-truth intersection divided by the sum of their areas.A perfect segmentation scores 1, while a completely mistaken segmentation scores 0.Source: B. Loss metric · p. 3
  • EddyNet uses a soft, differentiable Dice coefficient that retains softmax probabilities without binarization.The probabilities are compared with one-hot ground-truth class indicators.Source: B. Loss metric · p. 3
  • The multiclass loss averages three one-vs-all soft Dice coefficients, one for each class.The network minimizes one minus this soft mean Dice coefficient.Source: B. Loss metric · p. 3

A. Assessment of the performance

The evaluation compares activation functions and training losses on held-out SSH patches, finding that Dice loss improves relevant-class and overall Dice scores while ReLU+BN achieves better metrics at longer training cost.

A. Assessment of the performance

  • Example segmentation outputs label anticyclonic eddies green, cyclonic eddies brown, and non-eddy regions blue.The examples compare outputs from EddyNet and EddyNet S.Source: Figure 4 · p. 4, A. Assessment of the performance · p. 4
  • The study compares EddyNet and EddyNet S, along with Dice Loss and categorical cross-entropy, using test-set metrics from 50 random patch sets.Each set contains 360 SSH 120 × 120 maps from 2012.Source: A. Assessment of the performance · pp. 3–4, Table I · p. 4
  • Dice loss yields higher Dice coefficients for cyclonic and anticyclonic classes in both architectures and a better overall mean Dice coefficient than CCE loss.The comparison uses the original, non-soft mean Dice coefficient.Source: A. Assessment of the performance · p. 4
  • EddyNet with ReLU+BN obtains better metrics than EddyNet S with SELU, but requires a longer training procedure.Training EddyNet S takes nearly half the time needed for training EddyNet.Source: A. Assessment of the performance · p. 4

B. Ghost eddies

EddyNet is evaluated for identifying ghost eddies, which conventional detection and tracking algorithms may miss between consecutive maps.

B. Ghost eddies

  • EddyNet assigns ghost-eddy centers to the correct classes 55% of the time for anticyclonic eddies and 45% for cyclonic eddies.The authors present EddyNet as potentially relevant for detecting ghost eddies missed by conventional methods.Source: B. Ghost eddies · p. 4

VI. CONCLUSION

EddyNet is presented as a transferable deep-learning tool for oceanic eddy detection and classification, with extensions proposed for temporal, geographic, and multisource applications. The work also emphasizes practical reuse through training, fine-tuning, and integration with existing methods.

  • EddyNet applies deep-learning image-segmentation methods to eddy detection and classification from SSH maps.Source: §VI · p. 4
  • Future work includes temporal SSH volumes, a 3D architecture, additional surface information such as Sea Surface Temperature, and global deployment.Source: §VI · p. 4
  • The authors identify omitted postprocessing constraints and eddy tracking as further development opportunities.Source: §VI · p. 4
  • Users can train EddyNet from scratch with their own segmentation ground truth or fine-tune the provided weights on their datasets.Source: §VI · p. 4
  • Figure 5 compares marked ghost-eddy centers in SSH maps with PET14 and EddyNet segmentations using distinct class colors.Source: Figure 5 · p. 5

Related papers

Loading 1711.03954v1…