Successful training of end-to-end deep networks for real motion deblurring requires datasets of sharp/blurred image pairs that are realistic and diverse enough to achieve generalization to real blurred images. Obtaining such datasets remains a challenging task. In this paper, we first review the limitations of existing deblurring benchmark datasets from the perspective of generalization to blurry images in the wild. Secondly, we propose an efficient procedural methodology to generate sharp/blurred image pairs, based on a simple yet effective model for the formation of blurred images. This allows generating virtually unlimited realistic and diverse training pairs. We demonstrate the effectiveness of the proposed dataset by training existing deblurring architectures on the simulated pairs and evaluating them across four standard datasets of real blurred images. We observed superior generalization performance for the ultimate task of deblurring real motion-blurred photos of dynamic scenes when training with the proposed method.
翻译:成功培训真正运动变形的端到端深网络,需要现实和多样化的尖锐/粗糙图像配对数据集,这些数据集足以实现对真实模糊图像的概括化。 获取这些数据集仍是一项艰巨的任务。 在本文件中,我们首先从概括化角度审查现有模糊基准数据集的局限性,然后从荒野模糊图像的概括化角度加以审查。 其次,我们提出了一个高效的程序方法,以简单而有效的模型为基础,生成锐利/粗糙图像配对。 这样可以产生几乎无限的现实和多样的培训配对。 我们通过在模拟配对上培训现有脱泡结构,并通过对真实模糊图像的四种标准数据集进行评估,来展示拟议数据集的有效性。 我们观察到在用拟议方法培训时,在对动态场进行分流真实的图像进行分流的脱色照片的最终任务中,我们观察到了高级的普及性表现。