Prithvi-Complimentary Adaptive Fusion Encoder (CAFE): unlocking full-potential for flood inundation mapping

TL;DR

The paper introduces Prithvi-CAFE, a novel fusion encoder that combines GFMs with CNNs to significantly improve flood inundation mapping accuracy, especially in capturing local details.

Contribution

Prithvi-CAFE is a new model that integrates pretrained GFMs with CNN residual branches using attention modules for enhanced flood mapping.

Findings

Achieves state-of-the-art IoU scores on Sen1Flood11 and FloodPlanet datasets.

Outperforms baseline U-Net and other GFMs in flood inundation segmentation.

Effectively captures local details while maintaining long-range dependencies.

Abstract

Geo-Foundation Models (GFMs), have proven effective in diverse downstream applications, including semantic segmentation, classification, and regression tasks. However, in case of flood mapping using Sen1Flood11 dataset as a downstream task, GFMs struggles to outperform the baseline U-Net, highlighting model's limitation in capturing critical local nuances. To address this, we present the Prithvi-Complementary Adaptive Fusion Encoder (CAFE), which integrate Prithvi GFM pretrained encoder with a parallel CNN residual branch enhanced by Convolutional Attention Modules (CAM). Prithvi-CAFE enables fast and efficient fine-tuning through adapters in Prithvi and performs multi-scale, multi-level fusion with CNN features, capturing critical local details while preserving long-range dependencies. We achieve state-of-the-art results on two comprehensive flood mapping datasets: Sen1Flood11 and FloodPlanet. On Sen1Flood11 test data, Prithvi-CAFE (IoU 83.41) outperforms the original Prithvi (IoU 82.50) and other major GFMs (TerraMind 82.90, DOFA 81.54, spectralGPT: 81.02). The improvement is even more pronounced on the hold-out test site, where Prithvi-CAFE achieves an IoU of 81.37 compared to the baseline U-Net (70.57) and original Prithvi (72.42). On FloodPlanet, Prithvi-CAFE also surpasses the baseline U-Net and other GFMs, achieving an IoU of 64.70 compared to U-Net (60.14), Terramind (62.33), DOFA (59.15) and Prithvi 2.0 (61.91). Our proposed simple yet effective Prithvi-CAFE demonstrates strong potential for improving segmentation tasks where multi-channel and multi-modal data provide complementary information and local details are critical. The code is released on \href{https://github.com/Sk-2103/Prithvi-CAFE}{Prithvi-CAFE Github}