We present a novel adversarial distortion learning (ADL) for denoising two- and three-dimensional (2D/3D) biomedical image data. The proposed ADL consists of two auto-encoders: a denoiser and a discriminator. The denoiser removes noise from input data and the discriminator compares the denoised result to its noise-free counterpart. This process is repeated until the discriminator cannot differentiate the denoised data from the reference. Both the denoiser and the discriminator are built upon a proposed auto-encoder called Efficient-Unet. Efficient-Unet has a light architecture that uses the residual blocks and a novel pyramidal approach in the backbone to efficiently extract and re-use feature maps. During training, the textural information and contrast are controlled by two novel loss functions. The architecture of Efficient-Unet allows generalizing the proposed method to any sort of biomedical data. The 2D version of our network was trained on ImageNet and tested on biomedical datasets whose distribution is completely different from ImageNet; so, there is no need for re-training. Experimental results carried out on magnetic resonance imaging (MRI), dermatoscopy, electron microscopy and X-ray datasets show that the proposed method achieved the best on each benchmark. Our implementation and pre-trained models are available at https://github.com/mogvision/ADL.

翻译：我们提出了一种新颖的对抗性失真学习（ADL）方法，用于对二维和三维（2D/3D）生物医学图像数据进行去噪。所提出的ADL包含两个自编码器：一个去噪器和一个判别器。去噪器从输入数据中去除噪声，判别器则将去噪后的结果与无噪声的参考数据进行对比。该过程反复进行，直至判别器无法区分去噪数据与参考数据。去噪器和判别器均基于我们提出的一种名为Efficient-Unet的自编码器构建。Efficient-Unet采用轻量级架构，利用残差块和骨干网络中的新型金字塔方法高效提取并复用特征图。在训练过程中，通过两个新颖的损失函数控制纹理信息和对比度。Efficient-Unet的架构使得所提方法能够推广到任意类型的生物医学数据。我们的二维网络版本在ImageNet上训练，并在分布完全不同于ImageNet的生物医学数据集上测试，因此无需重新训练。在磁共振成像（MRI）、皮肤镜、电子显微镜和X射线数据集上进行的实验结果表明，所提方法在各项基准测试中均达到最佳性能。我们的实现和预训练模型可从https://github.com/mogvision/ADL获取。