DeblurGAN: Blind Motion Deblurring Using Conditional Adversarial Networks

By Orest Kupyn, Volodymyr Budzan, Mykola Mykhailych, Dmytro Mishkin, Jiri Matas. 2017

1,703 citations recorded by the source corpus.

We present DeblurGAN, an end-to-end learned method for motion deblurring. The learning is based on a conditional GAN and the content loss. DeblurGAN achieves state-of-the art performance both in the structural similarity measure and visual appearance. The quality of the deblurring model is also evaluated in a novel way on a real-world problem - object detection on (de-)blurred images. The method is 5 times faster than the closest competitor - Deep-Deblur [25]. We also introduce a novel method for generating synthetic motion blurred images from sharp ones, allowing realistic dataset augmentation. The model, code and the dataset are available at https://github.com/KupynOrest/DeblurGAN

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Citations in the corpus

AutoGAN: Neural Architecture Search for Generative Adversarial Networks (2019, background): “It has achieved great success in many generation and synthesis tasks, such as text-to-image translation [66, 65, 61, 48], image-to-image translation [22, 70, 63], and image enhancement [33, 31, 23].”
The Perception-Distortion Tradeoff (2017, background): “GAN-based image restoration Most recently, GAN-based methods have demonstrated unprecedented perceptual quality in super-resolution [3], [6], [9], [37], inpainting [7], [20], [38], compression [21], [39], [40], deblurring [41] and image-to-image translation [11], [22], [42].”
Deep Learning Techniques for Inverse Problems in Imaging (2020, background): “Another approach is DeblurGAN [92], in which a GAN is trained end-to-end using blurry images as input.”
Seeing Through Fog Without Seeing Fog: Deep Multimodal Sensor Fusion in Unseen Adverse Weather (2019, background): “Especially fog and haze removal from conventional intensity image data have been explored extensively [67, 70, 33, 53, 36, 7, 37, 46].”
Spatially-Attentive Patch-Hierarchical Network for Adaptive Motion Deblurring (2020, methodology): “We provide extensive comparisons with state-of-the-art learning-based methods, namely MS-CNN[13], DeblurGAN[8], DeblurGAN-v2[9], SRN[20], and Stack(4)-DMPHN[26].”
Multi-Stage Progressive Image Restoration (2021, methodology): “For example, residual learning [29] has been used to perform image denoising [70, 90], image deblurring [40, 41] and image deraining [36].”
TransGAN: Two Transformers Can Make One Strong GAN (2021, background): “GANs are widely adopted in image-to-image translation (Isola et al., 2017; Zhu et al., 2017a), image enhancement (Jiang et al., 2021; Ledig et al., 2017; Kupyn et al., 2018), and image editing (Ouyang et al., 2018; Yu et al., 2018).”
Blind Face Restoration via Deep Multi-scale Component Dictionaries (2020, methodology): “tion Along with the benets brought by deep CNNs, single image restoration has achieved great success in many tasks, e.g., image super-resolution [9,19,24,44, 46], denoising [13,38,42,43], deblurring [22,29,41], and compression artifact removal [8,10,12]. Due to the specic facial structure, there are”
One-Shot Free-View Neural Talking-Head Synthesis for Video Conferencing (2020, background): “[21], GANs have shown promising results in various areas [48], such as unconditional image synthesis [21, 23, 32, 33, 34, 49, 61], image translation [8, 12, 27, 29, 46, 47, 58, 67, 73, 84, 102, 103], text-to-image translation [62, 90, 95], image processing [17, 18, 28, 36, 39, 41, 42, 44, 72, 78, 89”
Towards Real-World Blind Face Restoration with Generative Facial Prior (2021, background): “Blind face restoration aims at recovering high-quality faces from the low-quality counterparts suffering from unknown degradation, such as low-resolution [14, 50, 9], noise [72], blur [40, 60], compression artifacts [13], etc.”
Deep Image Deblurring: A Survey (2022, methodology): “Kupyn et al. (2019) proposed DeblurGAN-v2 to restore sharp Cascaded Networks A cascaded network contains several modules, which are sequentially concatenated to construct a deeper structure.”
EDVR: Video Restoration With Enhanced Deformable Convolutional Networks (2019, background): “Image restoration tasks such as super-resolution (SR) [5, 20, 41, 18, 45, 51] and deblurring [27, 15, 38] have experienced significant improvements over the last few years thanks to deep learning.”
Single Image Deraining: A Comprehensive Benchmark Analysis (2019, background): “Such task-specific evaluation philosophy follows [34, 15].”
DeblurGAN-v2: Deblurring (Orders-of-Magnitude) Faster and Better (2019, methodology): “Similar to [21, 29], we introduce a direct skip connection from the input to the output, so that the learning focuses on the residue.”
Image Restoration with Mean-Reverting Stochastic Differential Equations (2023, methodology): “For comparison, we report four milestone deblur-ring approaches: DeepDeblur (Nah et al., 2017), Deblur-GAN (Kupyn et al., 2018), DeblurGAN-v2 (Kupyn et al., 2019), DBGAN (Zhang et al., 2020), and MAXIM (Tu et al., 2022).”
Deep Generalized Unfolding Networks for Image Restoration (2022, methodology): “For image deblurring, similar to [36, 55, 63, 79, 80], we train our model with 2, 103 image pairs from GoPro [47] dataset and evaluate each method on the test sets from GoPro and HIDE [58], which contain 1, 111 and 2, 025 samples, respectively.”
NTIRE 2019 Challenge on Video Deblurring and Super-Resolution: Dataset and Study (2019, background): “There have been several studies about deblurring and superresolution which aimed to improve perceptual quality with adversarial and perceptual losses [23, 17, 18, 10, 39].”
Artificial intelligence in the creative industries: a review (2020, methodology): “ generally unknown. Deblurring methods sharpen an image to increase subjective quality, and also to assist subsequent operations such as optical character recognition (OCR) [106] and object detection [113]. Early work in this area analysed the statistics of the image and attempted to model physical ”