Skip to main content
QUICK REVIEW

[论文解读] Exchange Is All You Need for Remote Sensing Change Detection

Sijun Dong, Siming Fu|arXiv (Cornell University)|Jan 12, 2026
Remote-Sensing Image Classification被引用 0
一句话总结

引入 SEED(Siamese Encoder–Exchange–Decoder),一种无参数的特征交换变更检测框架,替代显式差分,通过正交特征交换实现信息保持,在五个基准和三种骨干网络上取得竞争性或更优的结果。

ABSTRACT

Remote sensing change detection fundamentally relies on the effective fusion and discrimination of bi-temporal features. Prevailing paradigms typically utilize Siamese encoders bridged by explicit difference computation modules, such as subtraction or concatenation, to identify changes. In this work, we challenge this complexity with SEED (Siamese Encoder-Exchange-Decoder), a streamlined paradigm that replaces explicit differencing with parameter-free feature exchange. By sharing weights across both Siamese encoders and decoders, SEED effectively operates as a single parameter set model. Theoretically, we formalize feature exchange as an orthogonal permutation operator and prove that, under pixel consistency, this mechanism preserves mutual information and Bayes optimal risk, whereas common arithmetic fusion methods often introduce information loss. Extensive experiments across five benchmarks, including SYSU-CD, LEVIR-CD, PX-CLCD, WaterCD, and CDD, and three backbones, namely SwinT, EfficientNet, and ResNet, demonstrate that SEED matches or surpasses state of the art methods despite its simplicity. Furthermore, we reveal that standard semantic segmentation models can be transformed into competitive change detectors solely by inserting this exchange mechanism, referred to as SEG2CD. The proposed paradigm offers a robust, unified, and interpretable framework for change detection, demonstrating that simple feature exchange is sufficient for high performance information fusion. Code and full training and evaluation protocols will be released at https://github.com/dyzy41/open-rscd.

研究动机与目标

  • 通过消除双时相特征之间的显式差分计算,推动变更检测的简化;
  • 提出一种无参数的交换机制,保持信息并实现统一的编码器–解码器设计;
  • 从理论上证明特征交换是一种信息保持的置换,并在不同数据集和骨干网络上通过实证验证其有效性。

提出的方法

  • 提出 SEED:带权重共享的 Siamese 编码器和解码器,以及一个无参数的特征交换阶段;
  • 将特征交换形式化为一种正交置换算子,在像素一致性下保持互信息;
  • 评估三种交换方案(特征图层、通道和空间交换)及随机化变体,并给出 SEG2CD 配方将分割模型转换为变更检测器;
  • 使用单一的编码器–解码器参数集,不使用显式融合/差分模块,采用简单的逐层解码器和共享的 FPN 颈部结构;
  • 提供理论分析,显示在交换下互信息和 Bayes 最优风险不变性,与相加/相减融合带来的信息损失风险对比。

实验结果

研究问题

  • RQ1一个无参数的特征交换机制能否在不牺牲性能的前提下替代遥感变更检测中的显式差分?
  • RQ2通过正交置换交换双时相特征是否能保持信息并在像素一致性下支持有效的变更检测?
  • RQ3层级/通道/空间交换变体如何比较,且是否可以通过 SEED(SEG2CD)将分割模型直接转化为变更检测器?

主要发现

  • SEED 在五个基准(SYSU-CD、LEVIR-CD、PX-CLCD、WaterCD、CDD)上达到或超越最先进方法。
  • SEED 在三种骨干网络(SwinT、EfficientNet、ResNet)上实现竞争性结果。
  • 形式分析显示特征交换是一种信息保持的置换,在像素一致性下保持互信息和 Bayes 最优风险;像加法/减法之类非可逆融合可能导致信息损失。
  • 训练过程中的随机化交换充当置换正则化器,计算成本极低,通常性能稳定或有所提升。
  • SEG2CD 配方表明,标准分割模型通过仅插入交换即可转化为具有竞争力的变更检测器。

更好的研究,从现在开始

从论文设计到论文写作,大幅缩短您的研究时间。

无需绑定信用卡

本解读由 AI 生成,并经人工编辑审核。