Single Image Reflection Separation with Perceptual Losses

By Xuaner Cecilia Zhang, Ren Ng, Qifeng Chen. 2018

392 citations recorded by the source corpus.

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.

Subfields: Computer Vision and Pattern Recognition

How researchers see this paper

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

In Computer Vision and Pattern Recognition

Citations in the corpus

Speech Denoising with Deep Feature Losses (2018, background): “, image stylization, synthesis and reflection separation [25, 26, 27]).”
CelebA-Spoof: Large-Scale Face Anti-Spoofing Dataset with Rich Annotations (2020, methodology): “In detail, refer to [16], the ground truth of the depth map of live image is generated by PRNet [11] and the ground truth of the reflection map of the spoof image is generated by the method in [38].”
Portrait shadow manipulation (2020, methodology): “2019) and an ablation of our model with symmetry. in terms of PSNR, SSIM, and LPIPS on the “ facial-syn ” test dataset.”
“Double-DIP”: Unsupervised Image Decomposition via Coupled Deep-Image-Priors (2018, background): “Ambiguity can be resolved using external training, as in [40].”
Face Anti-Spoofing with Human Material Perception (2020, methodology): “The reﬂection maps are estimated by the state-of-the-art reﬂection estimation network [57] and then face regions are cropped from reﬂection maps to avoid being overﬁtted to backgrounds.”
Cooperative Semantic Segmentation and Image Restoration in Adverse Environmental Conditions (2019, methodology): “For image synthesis, we use two methods to synthesis reflection images as in [44, 52, 11] with some modifications to make sure the size of obtained images is identical with those in [8].”
Reflection Backdoor: A Natural Backdoor Attack on Deep Neural Networks (2020, methodology): “In our experiment, we adopt the state-of-the-art reflection removal method [74] to clean the poisoned data.”