In recent years, deep learning has driven remarkable progress in remote sensing image super resolution (RSISR). However, super-resolved images do not perform as well as original high-resolution images when directly used in remote sensing classification tasks, due to the loss of semantic information during the reconstruction process. To address this, we focus on semantic-oriented RSISR to improve the remote sensing classification performance using super-resolved images. We propose a dual-branch framework consisting of a semantic enhancement branch and an image reconstruction branch. The semantic enhancement branch exploits pretrained remote sensing classification networks to extract high-level semantic features, which are integrated into the reconstruction process. The reconstruction branch is built upon a novel hierarchical state space model, termed HierMamba, composed of hierarchical state space groups (HSSGs) to achieve multiscale and multigranularity feature extraction. By jointly leveraging semantic priors and HierMamba-based reconstruction, our method generates super-resolved images that are more favorable for downstream classification. Extensive experiments on RSISR benchmarks show that the proposed method produces superior visual quality and consistently improves classification performance compared with existing state-of-the-art methods.
Jinglei Hao, Yuting Lu, Shunzhou Wang et al.· IEEE Transactions on Geoscie...· 0 citations
Gaze Object Prediction (GOP) aims to localize and recognize the objects humans attend to, a task crucial for understanding human-centric interactions. However, existing methods are typically trained under a closed-vocabulary paradigm with a fixed label space and evaluated on scene-specific datasets, limiting their applicability to real-world scenarios where gaze targets often follow a long-tail distribution or belong to unseen categories. To address this gap, we introduce Diverse Scenes for Gaze object prediction (DiSG), a benchmark containing 86 in-the-wild categories that facilitates the evaluation of Open-Vocabulary GOP (OVGOP). Building on DiSG, we propose a framework that leverages text-driven object discovery to localize potential gaze candidates, with a gaze-guided selection module to pinpoint the intended target from the candidate objects. Furthermore, to better capture semantic knowledge across diverse in-the-wild categories, we introduce Gradient-Informed Selection Tuning (GIST) to selectively update parameters most relevant to a given class vocabulary. Extensive experiments demonstrate that our proposed model performs effectively in open-vocabulary settings and also outperforms existing methods in the conventional closed-vocabulary setting. The benchmark and code is available at https://github.com/sensniu/ovgop.
