Adaptive Denoising via GainTuning
Adaptive Denoising via GainTuning
复制标题
DOI：
--
复制
发表时间：
2021-07
期刊：
影响因子：
--
通讯作者：
S. Mohan;Joshua L. Vincent;R. Manzorro;P. Crozier;Eero P. Simoncelli;C. Fernandez‐Granda
中科院分区：
文献类型：
--
作者：
S. Mohan;Joshua L. Vincent;R. Manzorro;P. Crozier;Eero P. Simoncelli;C. Fernandez‐Granda
用于图像去噪的深度卷积神经网络（CNN）通常在大型数据集上进行训练。这些模型达到了当前的最先进水平，但是当应用于偏离训练分布的数据时，它们很难泛化。最近的工作表明，可以在单个噪声图像上训练降噪器。这些模型适应测试图像的特征，但它们的性能受到用于训练它们的少量信息的限制。在这里，我们提出“GainTuning”，其中在大型数据集上预训练的 CNN 模型针对各个测试图像进行自适应和选择性调整。为了避免过度拟合，GainTuning 优化 CNN 卷积层中每个通道的单个乘法缩放参数（“增益”）。我们表明，GainTuning 在标准图像去噪基准上改进了最先进的 CNN，提高了测试集中几乎所有图像的去噪性能。对于在噪声水平或图像类型上系统地不同于训练数据的测试图像，这些自适应改进甚至更加显着。我们展示了自适应降噪在科学应用中的潜力，其中 CNN 在合成数据上进行训练，并在真实的透射电子显微镜图像上进行测试。与现有方法相比，GainTuning 能够以极低的信噪比从这些数据忠实地重建催化纳米颗粒的结构。
Deep convolutional neural networks (CNNs) for image denoising are usually trained on large datasets. These models achieve the current state of the art, but they have difficulties generalizing when applied to data that deviate from the training distribution. Recent work has shown that it is possible to train denoisers on a single noisy image. These models adapt to the features of the test image, but their performance is limited by the small amount of information used to train them. Here we propose"GainTuning", in which CNN models pre-trained on large datasets are adaptively and selectively adjusted for individual test images. To avoid overfitting, GainTuning optimizes a single multiplicative scaling parameter (the"Gain") of each channel in the convolutional layers of the CNN. We show that GainTuning improves state-of-the-art CNNs on standard image-denoising benchmarks, boosting their denoising performance on nearly every image in a held-out test set. These adaptive improvements are even more substantial for test images differing systematically from the training data, either in noise level or image type. We illustrate the potential of adaptive denoising in a scientific application, in which a CNN is trained on synthetic data, and tested on real transmission-electron-microscope images. In contrast to the existing methodology, GainTuning is able to faithfully reconstruct the structure of catalytic nanoparticles from these data at extremely low signal-to-noise ratios.