警惕混叠——信号保留对鲁棒图像复原至关重要
Beware of Aliases -- Signal Preservation is Crucial for Robust Image Restoration
AI总结：
针对图像复原网络因混叠导致鲁棒性差的问题，提出BOA-Restormer，通过在频域执行部分下采样和上采样操作，确保无混叠路径，在低成本下提升模型鲁棒性。
摘要：
图像复原网络通常由编码器和解码器组成，分别负责从噪声、失真数据中聚合图像内容并恢复干净、无失真的图像。数据聚合以及高分辨率图像生成通常都伴随着混叠的风险，即标准架构为了在验证数据上达到高PSNR值而牺牲了重建模型输入的能力。代价是模型鲁棒性低。在这项工作中，我们表明，在先进的复原变换器中简单地提供无混叠路径，可以在低复原性能成本下支持改进的模型鲁棒性。为此，我们提出了BOA-Restormer，一种基于变换器的图像复原模型，它在频域中部分执行下采样和上采样操作，以确保整个模型的无混叠路径，同时可能保留所有相关的高频信息。
英文摘要：
Image restoration networks are usually comprised of an encoder and a decoder, responsible for aggregating image content from noisy, distorted data and to restore clean, undistorted images, respectively. Data aggregation as well as high-resolution image generation both usually come at the risk of involving aliases, i.e.~standard architectures put their ability to reconstruct the model input in jeopardy to reach high PSNR values on validation data. The price to be paid is low model robustness. In this work, we show that simply providing alias-free paths in state-of-the-art reconstruction transformers supports improved model robustness at low costs on the restoration performance. We do so by proposing BOA-Restormer, a transformer-based image restoration model that executes downsampling and upsampling operations partly in the frequency domain to ensure alias-free paths along the entire model while potentially preserving all relevant high-frequency information.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
