Separate In Latent Space: Unsupervised Single Image Layer Separation
This addresses the challenge of collecting aligned training data for vision tasks like intrinsic image decomposition, offering an incremental improvement over prior unsupervised approaches.
The paper tackles the ill-posed problem of single image layer separation, such as reflection removal, by proposing an unsupervised method that eliminates the need for ground truth triplet data, and it outperforms existing unsupervised methods in synthetic and real-world tasks.
Many real world vision tasks, such as reflection removal from a transparent surface and intrinsic image decomposition, can be modeled as single image layer separation. However, this problem is highly ill-posed, requiring accurately aligned and hard to collect triplet data to train the CNN models. To address this problem, this paper proposes an unsupervised method that requires no ground truth data triplet in training. At the core of the method are two assumptions about data distributions in the latent spaces of different layers, based on which a novel unsupervised layer separation pipeline can be derived. Then the method can be constructed based on the GANs framework with self-supervision and cycle consistency constraints, etc. Experimental results demonstrate its successfulness in outperforming existing unsupervised methods in both synthetic and real world tasks. The method also shows its ability to solve a more challenging multi-layer separation task.