Video restoration concerns the recovery of a clean video sequence starting from its degraded version. Different video restoration tasks exist, including denoising, deblurring, super-resolution, and reduction of compression artifacts. In this paper, we provide a comprehensive review of the main features of existing video restoration methods based on deep learning. We focus our attention on the main architectural components, strategies for motion handling, and loss functions. We analyze the standard benchmark datasets and use them to summarize the performance of video restoration methods, both in terms of effectiveness and efficiency. In conclusion, the main challenges and future research directions in video restoration using deep learning are highlighted.
Low-light video enhancement (LLVE) has received little attention compared to low-light image enhancement (LLIE) mainly due to the lack of paired low-/normal-light video datasets. Consequently, a common approach to LLVE is to enhance each video frame individually using LLIE methods. However, this practice introduces temporal inconsistencies in the resulting video. In this work, we propose a recurrent neural network (RNN) that, given a low-light video and its per-frame enhanced version, produces a temporally consistent video preserving the underlying frame-based enhancement. We achieve this by training our network with a combination of a new forwardbackward temporal consistency loss and a content-preserving loss. At inference time, we can use our trained network to correct videos processed by any LLIE method. Experimental results show that our method achieves the best trade-off between temporal consistency improvement and fidelity with the perframe enhanced video, exhibiting a lower memory complexity and comparable time complexity with respect to other state-ofthe-art methods for temporal consistency
Video restoration concerns the recovery of a clean video sequence starting from its degraded version. Different video restoration tasks exist, including denoising, deblurring, super-resolution, and reduction of compression artifacts. In this paper, we provide a comprehensive review of the main features of existing video restoration methods based on deep learning. We focus our attention on the main architectural components , strategies for motion handling, and loss functions. We analyze the standard benchmark datasets and use them to summarize the performance of video restoration methods, both in terms of effectiveness and efficiency. In conclusion, the main challenges and future research directions in video restoration using deep learning are highlighted.
Accurate building segmentation from high-resolution aerial imagery is essential for numerous applications in remote sensing, urban planning, and disaster management. While AI-based methods enable fast, scalable, and cost-effective segmentation of building footprints, their development is often limited by the scarce availability of large-scale, geographically diverse datasets with reliable pixel-level annotations. In this work, we present SegFVG, a large-scale, high-resolution, and geographically diverse dataset for building segmentation, focused on the Friuli Venezia Giulia region in northeastern Italy. The dataset includes over 15,000 true orthophoto aerial image tiles, each of size 2000 × 2000 pixels with a ground sampling distance of 0.1 meters, paired with precise pixel-level building segmentation masks. Covering approximately 616 km 2 , SegFVG captures a broad spectrum of urban, suburban, and rural settings across varied landscapes, including mountainous, flat, and coastal areas. Alongside the dataset, we provide benchmark results using several deep learning models. These support the usability of SegFVG for the development of accurate segmentation models and serve as a baseline to accelerate future research in building segmentation.