Automatic Reflection Removal using Gradient Intensity and Motion Cues

Sun, Chao; Liu, Shuaicheng; Yang, Taotao; Zeng, Bing; Wang, Zhengning; Liu, Guanghui

doi:10.1145/2964284.2967264

Cited by 56 publications

(27 citation statements)

References 15 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…As the reflection separation problem is ill-posed, most previous work tackles this problem with multiple input images. These multi-image approaches often use motion cues to separate the transmission and reflection layers [32,9,20,28,23,6,29,10]. The motion cues are either inferred from calibrated cameras, or motion parallax that assumes the background and reflection objects have greatly different motion fields.…”

Section: Related Workmentioning

confidence: 99%

Single Image Reflection Separation with Perceptual Losses

Zhang

Chen

2018

2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition

280

409

View full text Add to dashboard Cite

We present an approach to separating reflection from a single image. The approach uses a fully convolutional network trained end-to-end with losses that exploit low-level and high-level image information. Our loss function includes two perceptual losses: a feature loss from a visual perception network, and an adversarial loss that encodes characteristics of images in the transmission layers. We also propose a novel exclusion loss that enforces pixel-level layer separation. We create a dataset of real-world images with reflection and corresponding ground-truth transmission layers for quantitative evaluation and model training. We validate our method through comprehensive quantitative experiments and show that our approach outperforms state-of-the-art reflection removal methods in PSNR, SSIM, and perceptual user study. We also extend our method to two other image enhancement tasks to demonstrate the generality of our approach.

show abstract

Section: Related Workmentioning

confidence: 99%

Single Image Reflection Separation with Perceptual Losses

Zhang

Chen

2018

2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition

280

409

View full text Add to dashboard Cite

show abstract

“…Guo et al [6] use rank constraints and structural priors to exploit the correlation of the transmission layers from multiple images. In [11,20] the authors use SIFT-flow to calculate the motion from photos taken from different view-points. Using a motion score they classify edges as belonging to either transmission or reflection layers, which helps solve an optimization scheme to separate the layers.…”

Section: Multiple-image Reflection Removalmentioning

confidence: 99%

“…For example, professional photographers use polarizing filters to mitigate, if not eliminate, reflection artifacts. Similarly, using multiple photographs taken with small displacements makes the problem easier to tackle [5,6,11,14,18,20,21,24]. However, given the skill and resources of an average user, none of these methods are feasible.…”

Section: Introductionmentioning

confidence: 99%

Single Image Reflection Suppression

Arvanitopoulos

Achanta

Süsstrunk

2017

2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)

122

120

View full text Add to dashboard Cite

Reflections are a common artifact in images taken through glass windows. Automatically removing the reflection artifacts after the picture is taken is an ill-posed problem. Attempts to solve this problem using optimization schemes therefore rely on various prior assumptions from the physical world. Instead of removing reflections from a single image, which has met with limited success so far, we propose a novel approach to suppress reflections. It is based on a Laplacian data fidelity term and an l 0 gradient sparsity term imposed on the output. With experiments on artificial and real-world images we show that our reflection suppression method performs better than the state-of-theart reflection removal techniques.

show abstract

“…Schechner et al [21] separate the layers by using a pair of images captured with different focus settings. Others separate the layers by exploiting motion cues and minimizing the layer correlation in a short video sequence [4,7,8,10,17,18,20,27,28,29,31].…”

Section: Prior Workmentioning

confidence: 99%

“…This is particularly evident for the assignment of the average color of each layer (see Figure 2). To address these ambiguities, previous methods seek to get additional information either through a custom hardware design [13,14,22,23] or multiple images [4,7,8,10,17,20,27,28,29,31]. However, reflection removal using a single image remains a very appealing and useful problem, which is gaining attention in the scientific community [2,5,15,16].…”

Section: Introductionmentioning

confidence: 99%

Learning to see through reflections

Jin

Süsstrunk

Favaro

2018

2018 IEEE International Conference on Computational Photography (ICCP)

View full text Add to dashboard Cite

Pictures of objects behind a glass are difficult to interpret and understand due to the superposition of two real images: a reflection layer and a background layer. Separation of these two layers is challenging due to the ambiguities in assigning texture patterns and the average color in the input image to one of the two layers. In this paper, we propose a novel method to reconstruct these layers given a single input image by explicitly handling the ambiguities of the reconstruction. Our approach combines the ability of neural networks to build image priors on large image regions with an image model that accounts for the brightness ambiguity and saturation. We find that our solution generalizes to real images even in the presence of strong reflections. Extensive quantitative and qualitative experimental evaluations on both real and synthetic data show the benefits of our approach over prior work. Moreover, our proposed neural network is computationally and memory efficient.

show abstract

Automatic Reflection Removal using Gradient Intensity and Motion Cues

Cited by 56 publications

References 15 publications

Single Image Reflection Separation with Perceptual Losses

Single Image Reflection Separation with Perceptual Losses

Single Image Reflection Suppression

Learning to see through reflections

Contact Info

Product

Resources

About