Deblurring Face Images using Uncertainty Guided Multi-Stream Semantic Networks
This addresses the problem of enhancing blurry face images for applications like surveillance or photography, but it is incremental as it builds on existing semantic segmentation and deblurring techniques.
The paper tackles face image deblurring by proposing a multi-stream architecture that uses semantic labels and uncertainty guidance, achieving significant improvements over state-of-the-art methods on three datasets.
We propose a novel multi-stream architecture and training methodology that exploits semantic labels for facial image deblurring. The proposed Uncertainty Guided Multi- Stream Semantic Network (UMSN) processes regions belonging to each semantic class independently and learns to combine their outputs into the final deblurred result. Pixel-wise semantic labels are obtained using a segmentation network. A predicted confidence measure is used during training to guide the network towards the challenging regions of the human face such as the eyes and nose. The entire network is trained in an end- to-end fashion. Comprehensive experiments on three different face datasets demonstrate that the proposed method achieves significant improvements over the recent state-of-the-art face deblurring methods. Code is available at: https://github.com/ rajeevyasarla/UMSN-Face-Deblurring