NGram-MoSE: Efficient Remote Sensing Super-Resolution via N-Gram Context and Mixture-of-Experts

2026-06-07

AI Summary

This work addresses the challenge in remote sensing image super-resolution of simultaneously achieving computational efficiency and robust out-of-distribution generalization, where performance typically degrades significantly on unseen geographic regions. The authors propose a lightweight Transformer architecture that innovatively incorporates N-Gram context injection to enhance local texture continuity and mitigate window boundary artifacts. This design is further augmented with a sparsely activated Mixture-of-Experts (MoE) feed-forward network, which increases model capacity with negligible additional inference cost. Evaluated on out-of-distribution test sets, the method achieves a PSNR of 31.68 dB while reducing FLOPs by 14× compared to existing approaches. When applied to a downstream landslide segmentation task, it improves mAP@50 by 4.47%, substantially outperforming the bicubic interpolation baseline.

Abstract

Remote sensing applications for environmental monitoring and disaster management are frequently constrained by a spatial--temporal trade-off: imagery with fine spatial detail is often acquired less frequently, whereas more temporally available observations are typically coarser. Single-image super-resolution provides a practical means to enhance coarse imagery without changing acquisition schedules, yet many Transformer-based SR models remain computationally expensive and can be sensitive to limited or geographically biased training data, which degrades robustness under out-of-distribution conditions. This paper presents NGram-MoSE, a lightweight Transformer architecture designed to improve both efficiency and texture continuity. NGram-MoSE introduces N-Gram Context Injection to strengthen cross-window local consistency and mitigate window-boundary artifacts, and incorporates a Mixture-of-Experts (MoE) feed-forward design to scale capacity through sparse activation without proportional growth in inference cost. Experiments on a geographically disjoint OOD test set show that NGram-MoSE achieves 31.68\,dB PSNR while reducing FLOPs by \(14\times\) relative to a heavyweight Transformer reference. Downstream evaluation on a landslide segmentation benchmark further demonstrates that restoring degraded inputs to the detector training scale improves performance, yielding a 4.47\% absolute gain in mAP@50 over bicubic upsampling, and exhibits stronger cross-scale consistency under scale extrapolation. These results indicate that NGram-MoSE provides an effective SR module for resource-constrained remote sensing pipelines requiring robust generalization.

Problem

• super-resolution
• remote sensing
• out-of-distribution
• computational efficiency
• texture continuity

Innovation

• N-Gram Context Injection
• Mixture-of-Experts
• Lightweight Transformer
• Super-Resolution
• Out-of-Distribution Generalization