Temporal modeling is crucial for video super-resolution. Most of the video super-resolution methods adopt the optical flow or deformable convolution for explicitly motion compensation. However, such temporal modeling techniques increase the model complexity and might fail in case of occlusion or complex motion, resulting in serious distortion and artifacts. In this paper, we propose to explore the role of explicit temporal difference modeling in both LR and HR space. Instead of directly feeding consecutive frames into a VSR model, we propose to compute the temporal difference between frames and divide those pixels into two subsets according to the level of difference. They are separately processed with two branches of different receptive fields in order to better extract complementary information. To further enhance the super-resolution result, not only spatial residual features are extracted, but the difference between consecutive frames in high-frequency domain is also computed. It allows the model to exploit intermediate SR results in both future and past to refine the current SR output. The difference at different time steps could be cached such that information from further distance in time could be propagated to the current frame for refinement. Experiments on several video super-resolution benchmark datasets demonstrate the effectiveness of the proposed method and its favorable performance against state-of-the-art methods.
Human actions involve complex pose variations and their 2D projections can be highly ambiguous. Thus 3D spatio-temporal or 4D (i.e., 3D+T) human skeletons, which are photometric and viewpoint invariant, are an excellent alternative to 2D+T skeletons/pixels to improve action recognition accuracy. This paper proposes a new 4D dataset HAA4D which consists of more than 3,300 RGB videos in 300 human atomic action classes. HAA4D is clean, diverse, class-balanced where each class is viewpoint-balanced with the use of 4D skeletons, in which as few as one 4D skeleton per class is sufficient for training a deep recognition model. Further, the choice of atomic actions makes annotation even easier, because each video clip lasts for only a few seconds. All training and testing 3D skeletons in HAA4D are globally aligned, using a deep alignment model to the same global space, making each skeleton face the negative z-direction. Such alignment makes matching skeletons more stable by reducing intraclass variations and thus with fewer training samples per class needed for action recognition. Given the high diversity and skeletal alignment in HAA4D, we construct the first baseline few-shot 4D human atomic action recognition network without bells and whistles, which produces comparable or higher performance than relevant state-of-the-art techniques relying on embedded space encoding without explicit skeletal alignment, using the same small number of training samples of unseen classes.
In most video platforms, such as Youtube, and TikTok, the played videos usually have undergone multiple video encodings such as hardware encoding by recording devices, software encoding by video editing apps, and single/multiple video transcoding by video application servers. Previous works in compressed video restoration typically assume the compression artifacts are caused by one-time encoding. Thus, the derived solution usually does not work very well in practice. In this paper, we propose a new method, temporal spatial auxiliary network (TSAN), for transcoded video restoration. Our method considers the unique traits between video encoding and transcoding, and we consider the initial shallow encoded videos as the intermediate labels to assist the network to conduct self-supervised attention training. In addition, we employ adjacent multi-frame information and propose the temporal deformable alignment and pyramidal spatial fusion for transcoded video restoration. The experimental results demonstrate that the performance of the proposed method is superior to that of the previous techniques. The code is available at https://github.com/icecherylXuli/TSAN.
We present NeRF-SR, a solution for high-resolution (HR) novel view synthesis with mostly low-resolution (LR) inputs. Our method is built upon Neural Radiance Fields (NeRF) that predicts per-point density and color with a multi-layer perceptron. While producing images at arbitrary scales, NeRF struggles with resolutions that go beyond observed images. Our key insight is that NeRF has a local prior, which means predictions of a 3D point can be propagated in the nearby region and remain accurate. We first exploit it by a super-sampling strategy that shoots multiple rays at each image pixel, which enforces multi-view constraint at a sub-pixel level. Then, we show that NeRF-SR can further boost the performance of super-sampling by a refinement network that leverages the estimated depth at hand to hallucinate details from related patches on an HR reference image. Experiment results demonstrate that NeRF-SR generates high-quality results for novel view synthesis at HR on both synthetic and real-world datasets.
Two-stage and query-based instance segmentation methods have achieved remarkable results. However, their segmented masks are still very coarse. In this paper, we present Mask Transfiner for high-quality and efficient instance segmentation. Instead of operating on regular dense tensors, our Mask Transfiner decomposes and represents the image regions as a quadtree. Our transformer-based approach only processes detected error-prone tree nodes and self-corrects their errors in parallel. While these sparse pixels only constitute a small proportion of the total number, they are critical to the final mask quality. This allows Mask Transfiner to predict highly accurate instance masks, at a low computational cost. Extensive experiments demonstrate that Mask Transfiner outperforms current instance segmentation methods on three popular benchmarks, significantly improving both two-stage and query-based frameworks by a large margin of +3.0 mask AP on COCO and BDD100K, and +6.6 boundary AP on Cityscapes. Our code and trained models will be available at http://vis.xyz/pub/transfiner.
Conventional video inpainting is neither object-oriented nor occlusion-aware, making it liable to obvious artifacts when large occluded object regions are inpainted. This paper presents occlusion-aware video object inpainting, which recovers both the complete shape and appearance for occluded objects in videos given their visible mask segmentation. To facilitate this new research, we construct the first large-scale video object inpainting benchmark YouTube-VOI to provide realistic occlusion scenarios with both occluded and visible object masks available. Our technical contribution VOIN jointly performs video object shape completion and occluded texture generation. In particular, the shape completion module models long-range object coherence while the flow completion module recovers accurate flow with sharp motion boundary, for propagating temporally-consistent texture to the same moving object across frames. For more realistic results, VOIN is optimized using both T-PatchGAN and a new spatio-temporal attention-based multi-class discriminator. Finally, we compare VOIN and strong baselines on YouTube-VOI. Experimental results clearly demonstrate the efficacy of our method including inpainting complex and dynamic objects. VOIN degrades gracefully with inaccurate input visible mask.
Multiple object tracking and segmentation requires detecting, tracking, and segmenting objects belonging to a set of given classes. Most approaches only exploit the temporal dimension to address the association problem, while relying on single frame predictions for the segmentation mask itself. We propose Prototypical Cross-Attention Network (PCAN), capable of leveraging rich spatio-temporal information for online multiple object tracking and segmentation. PCAN first distills a space-time memory into a set of prototypes and then employs cross-attention to retrieve rich information from the past frames. To segment each object, PCAN adopts a prototypical appearance module to learn a set of contrastive foreground and background prototypes, which are then propagated over time. Extensive experiments demonstrate that PCAN outperforms current video instance tracking and segmentation competition winners on both Youtube-VIS and BDD100K datasets, and shows efficacy to both one-stage and two-stage segmentation frameworks. Code will be available at http://vis.xyz/pub/pcan.
This paper presents a simple yet effective approach to modeling space-time correspondences in the context of video object segmentation. Unlike most existing approaches, we establish correspondences directly between frames without re-encoding the mask features for every object, leading to a highly efficient and robust framework. With the correspondences, every node in the current query frame is inferred by aggregating features from the past in an associative fashion. We cast the aggregation process as a voting problem and find that the existing inner-product affinity leads to poor use of memory with a small (fixed) subset of memory nodes dominating the votes, regardless of the query. In light of this phenomenon, we propose using the negative squared Euclidean distance instead to compute the affinities. We validated that every memory node now has a chance to contribute, and experimentally showed that such diversified voting is beneficial to both memory efficiency and inference accuracy. The synergy of correspondence networks and diversified voting works exceedingly well, achieves new state-of-the-art results on both DAVIS and YouTubeVOS datasets while running significantly faster at 20+ FPS for multiple objects without bells and whistles.
We introduce Few-Shot Video Object Detection (FSVOD) with three important contributions: 1) a large-scale video dataset FSVOD-500 comprising of 500 classes with class-balanced videos in each category for few-shot learning; 2) a novel Tube Proposal Network (TPN) to generate high-quality video tube proposals to aggregate feature representation for the target video object; 3) a strategically improved Temporal Matching Network (TMN+) to match representative query tube features and supports with better discriminative ability. Our TPN and TMN+ are jointly and end-to-end trained. Extensive experiments demonstrate that our method produces significantly better detection results on two few-shot video object detection datasets compared to image-based methods and other naive video-based extensions. Codes and datasets will be released at https://github.com/fanq15/FewX.
Despite the significant progress made by deep learning in natural image matting, there has been so far no representative work on deep learning for video matting due to the inherent technical challenges in reasoning temporal domain and lack of large-scale video matting datasets. In this paper, we propose a deep learning-based video matting framework which employs a novel and effective spatio-temporal feature aggregation module (ST-FAM). As optical flow estimation can be very unreliable within matting regions, ST-FAM is designed to effectively align and aggregate information across different spatial scales and temporal frames within the network decoder. To eliminate frame-by-frame trimap annotations, a lightweight interactive trimap propagation network is also introduced. The other contribution consists of a large-scale video matting dataset with groundtruth alpha mattes for quantitative evaluation and real-world high-resolution videos with trimaps for qualitative evaluation. Quantitative and qualitative experimental results show that our framework significantly outperforms conventional video matting and deep image matting methods applied to video in presence of multi-frame temporal information.