Image relighting, which involves modifying the lighting conditions while preserving the visual content, is fundamental to computer vision. This study introduced a bi-modal lightweight deep learning model for depth-guided relighting. The model utilizes the Res2Net Squeezed block’s ability to capture long-range dependencies and to enhance feature representation for both the input image and its corresponding depth map. The proposed model adopts an encoder–decoder structure with Res2Net Squeezed blocks integrated at each stage of encoding and decoding. The model was trained and evaluated on the VIDIT dataset, which consists of 300 triplets of images. Each triplet contains the input image, its corresponding depth map, and the relit image under diverse lighting conditions, such as different illuminant angles and color temperatures. The enhanced feature representation and improved information flow within the Res2Net Squeezed blocks enable the model to handle complex lighting variations and generate realistic relit images. The experimental results demonstrated the proposed approach’s effectiveness in relighting accuracy, measured by metrics such as the PSNR, SSIM, and visual quality.
Social media filters cause various manipulations in the original image. While these manipulations make the images visibly pleasant, they can affect their authenticity and the performance of further image-processing tasks. Therefore, removing these social media filters is a crucial pre-processing step for the data pipeline of any image-processing task. In this paper, a recurrent residual UNet-based model is proposed to remove the filters. The recurrent residual block allows the units of each layer to evolve over discrete time steps so that the activity of each unit is modulated by the activities of its neighboring units. Because of this characteristic, the networks learn contextual information more efficiently and effectively. Assuming that a filter can be noise, the proposed model learns the feature representation of the filter, and then the original, unfiltered image is retrieved by taking the residuals of the input image and the filter representation. The experimental results show that the proposed model outperforms both supervised and unsupervised methods by a considerable margin in the IFFI dataset.
This paper reviews the NTIRE 2020 challenge on real image denoising with focus on the newly introduced dataset, the proposed methods and their results. The challenge is a new version of the previous NTIRE 2019 challenge on real image denoising that was based on the SIDD benchmark. This challenge is based on a newly collected validation and testing image datasets, and hence, named SIDD+. This challenge has two tracks for quantitatively evaluating image denoising performance in (1) the Bayer-pattern rawRGB and (2) the standard RGB (sRGB) color spaces. Each track ~250 registered participants. A total of 22 teams, proposing 24 methods, competed in the final phase of the challenge. The proposed methods by the participating teams represent the current state-of-the-art performance in image denoising targeting real noisy images. The newly collected SIDD+ datasets are publicly available at: https://bit.ly/siddplus_data.