ResAttUNet: Detecting Marine Debris using an Attention activated Residual UNet

TL;DR

This paper introduces ResAttUNet, an attention-based residual UNet model that improves marine debris detection in remote sensing images, leveraging a novel spatial-aware architecture to outperform existing methods on the MARIDA dataset.

Contribution

The paper presents a novel attention-activated residual UNet architecture with a spatial-aware encoder-decoder design for enhanced marine debris segmentation.

Findings

Outperforms existing state-of-the-art on MARIDA dataset

Maintains contextual information in sparse ground truth patches

Provides open-source code for reproducibility

Abstract

Currently, a significant amount of research has been done in field of Remote Sensing with the use of deep learning techniques. The introduction of Marine Debris Archive (MARIDA), an open-source dataset with benchmark results, for marine debris detection opened new pathways to use deep learning techniques for the task of debris detection and segmentation. This paper introduces a novel attention based segmentation technique that outperforms the existing state-of-the-art results introduced with MARIDA. The paper presents a novel spatial aware encoder and decoder architecture to maintain the contextual information and structure of sparse ground truth patches present in the images. The attained results are expected to pave the path for further research involving deep learning using remote sensing images. The code is available at https://github.com/sheikhazhanmohammed/SADMA.git