Single Image Reflection Removal through Cascaded Refinement

Abstract

We address the problem of removing undesirable reflections from a single image captured through a glass surface, which is an ill-posed, challenging but practically important problem for photo enhancement. Inspired by iterative structure reduction for hidden community detection in social networks, we propose an Iterative Boost Convolutional LSTM Network (IBCLN) that enables cascaded prediction for reflection removal. IBCLN iteratively refines estimates of the transmission and reflection layers at each step in a manner that they can boost the prediction quality for each other. The intuition is that progressive refinement of the transmission or reflection layer is aided by increasingly better estimates of these quantities as input, and that transmission and reflection are complementary to each other in a single image and thus provide helpful auxiliary information for each other's prediction. To facilitate training over multiple cascade steps, we employ LSTM to address the vanishing gradient problem, and incorporate a reconstruction loss as further training guidance at each step. In addition, we create a dataset of real-world images with reflection and ground-truth transmission layers to mitigate the problem of insufficient data. Through comprehensive experiments, IB-CLN demonstrates performance that surpasses state-of-theart reflection removal methods.

Deep-Masking Generative Network: A Unified Framework for Background Restoration from Superimposed Images

“…The first type is prior-based method, which restores background image by referring to physical property in degradation such as edge blurring [22], [23], [24], streak directionality [25], [26], [27], or atmosphere scattering modeling [28], [9]. The second type of generic methods perform background restoration in a recursive manner [10], [11], [12], [13]. It recursively employ itself and performs supervision in each iteration, aiming to progressively refining background image.…”

“…It is expected to model the mapping function from the noise image to background image in a brute-force way by CNNs, which turns out to be hardly achieved due to the complicated noise patterns. An effective way to address this limitation is to recursively employ the same CNN-based module and improve the quality of the synthesized background image recurrently [10], [11], [12], [13]. While such method indeed boosts the performance of image background restoration to some degree, it is at the expense of more computation and consuming time.…”