Abstract: Here, we propose a hybrid Deep Learning (DL) framework consisting of a Denoising Autoencoder (DAE), Convolutional Neural Network (CNN), Bidirectional LSTM (BiLSTM), and a custom Attention ...

Abstract: Polar sea ice monitoring is critical for global climate research and Arctic ship navigation, yet traditional satellite remote sensing struggles to resolve microscale ice features due to ...