DNN for fish tracking and segmentation based on high-quality labels is expensive. Alternative unsupervised approaches rely on spatial and temporal variations that naturally occur in video data to generate noisy pseudo-ground-truth labels. These pseudo-labels are used to train a multi-task deep neural network. In this paper, we propose a three-stage framework for robust fish tracking and segmentation, where the first stage is an optical flow model, which generates the pseudo labels using spatial and temporal consistency between frames. In the second stage, a self-supervised model refines the pseudo-labels incrementally. In the third stage, the refined labels are used to train a segmentation network. No human annotations are used during the training or inference. Extensive experiments are performed to validate our method on three public underwater video datasets and to demonstrate that it is highly effective for video annotation and segmentation. We also evaluate the robustness of our framework to different imaging conditions and discuss the limitations of our current implementation.
Video frame interpolation is a challenging task due to the ever-changing real-world scene. Previous methods often calculate the bi-directional optical flows and then predict the intermediate optical flows under the linear motion assumptions, leading to isotropic intermediate flow generation. Follow-up research obtained anisotropic adjustment through estimated higher-order motion information with extra frames. Based on the motion assumptions, their methods are hard to model the complicated motion in real scenes. In this paper, we propose an end-to-end training method A^2OF for video frame interpolation with event-driven Anisotropic Adjustment of Optical Flows. Specifically, we use events to generate optical flow distribution masks for the intermediate optical flow, which can model the complicated motion between two frames. Our proposed method outperforms the previous methods in video frame interpolation, taking supervised event-based video interpolation to a higher stage.
Optical flow estimation in omnidirectional videos faces two significant issues: the lack of benchmark datasets and the challenge of adapting perspective video-based methods to accommodate the omnidirectional nature. This paper proposes the first perceptually natural-synthetic omnidirectional benchmark dataset with a 360-degree field of view, FLOW360, with 40 different videos and 4,000 video frames. We conduct comprehensive characteristic analysis and comparisons between our dataset and existing optical flow datasets, which manifest perceptual realism, uniqueness, and diversity. To accommodate the omnidirectional nature, we present a novel Siamese representation Learning framework for Omnidirectional Flow (SLOF). We train our network in a contrastive manner with a hybrid loss function that combines contrastive loss and optical flow loss. Extensive experiments verify the proposed framework's effectiveness and show up to 40% performance improvement over the state-of-the-art approaches. Our FLOW360 dataset and code are available at https://siamlof.github.io/.
Learned video compression methods have gained a variety of interest in the video coding community since they have matched or even exceeded the rate-distortion (RD) performance of traditional video codecs. However, many current learning-based methods are dedicated to utilizing short-range temporal information, thus limiting their performance. In this paper, we focus on exploiting the unique characteristics of video content and further exploring temporal information to enhance compression performance. Specifically, for long-range temporal information exploitation, we propose temporal prior that can update continuously within the group of pictures (GOP) during inference. In that case temporal prior contains valuable temporal information of all decoded images within the current GOP. As for short-range temporal information, we propose a progressive guided motion compensation to achieve robust and effective compensation. In detail, we design a hierarchical structure to achieve multi-scale compensation. More importantly, we use optical flow guidance to generate pixel offsets between feature maps at each scale, and the compensation results at each scale will be used to guide the following scale's compensation. Sufficient experimental results demonstrate that our method can obtain better RD performance than state-of-the-art video compression approaches. The code is publicly available on: https://github.com/Huairui/LSTVC.
This paper presents a state-of-the-art overview on how to architect, design, and optimize Deep Neural Networks (DNNs) such that performance is improved and accuracy is preserved. The paper covers a set of optimizations that span the entire Machine Learning processing pipeline. We introduce two types of optimizations. The first alters the DNN model and requires NN re-training, while the second does not. We focus on GPU optimizations, but we believe the presented techniques can be used with other AI inference platforms. To demonstrate the DNN model optimizations, we improve one of the most advanced deep network architectures for optical flow, RAFT arXiv:2003.12039, on a popular edge AI inference platform (Nvidia Jetson AGX Xavier).
Dense depth and pose estimation is a vital prerequisite for various video applications. Traditional solutions suffer from the robustness of sparse feature tracking and insufficient camera baselines in videos. Therefore, recent methods utilize learning-based optical flow and depth prior to estimate dense depth. However, previous works require heavy computation time or yield sub-optimal depth results. We present GCVD, a globally consistent method for learning-based video structure from motion (SfM) in this paper. GCVD integrates a compact pose graph into the CNN-based optimization to achieve globally consistent estimation from an effective keyframe selection mechanism. It can improve the robustness of learning-based methods with flow-guided keyframes and well-established depth prior. Experimental results show that GCVD outperforms the state-of-the-art methods on both depth and pose estimation. Besides, the runtime experiments reveal that it provides strong efficiency in both short- and long-term videos with global consistency provided.
Unsupervised optical flow estimators based on deep learning have attracted increasing attention due to the cost and difficulty of annotating for ground truth. Although performance measured by average End-Point Error (EPE) has improved over the years, flow estimates are still poorer along motion boundaries (MBs), where the flow is not smooth, as is typically assumed, and where features computed by neural networks are contaminated by multiple motions. To improve flow in the unsupervised settings, we design a framework that detects MBs by analyzing visual changes along boundary candidates and replaces motions close to detections with motions farther away. Our proposed algorithm detects boundaries more accurately than a baseline method with the same inputs and can improve estimates from any flow predictor without additional training.
Domain shifts, such as appearance changes, are a key challenge in real-world applications of activity recognition models, which range from assistive robotics and smart homes to driver observation in intelligent vehicles. For example, while simulations are an excellent way of economical data collection, a Synthetic-to-Real domain shift leads to a > 60% drop in accuracy when recognizing activities of Daily Living (ADLs). We tackle this challenge and introduce an activity domain generation framework which creates novel ADL appearances (novel domains) from different existing activity modalities (source domains) inferred from video training data. Our framework computes human poses, heatmaps of body joints, and optical flow maps and uses them alongside the original RGB videos to learn the essence of source domains in order to generate completely new ADL domains. The model is optimized by maximizing the distance between the existing source appearances and the generated novel appearances while ensuring that the semantics of an activity is preserved through an additional classification loss. While source data multimodality is an important concept in this design, our setup does not rely on multi-sensor setups, (i.e., all source modalities are inferred from a single video only.) The newly created activity domains are then integrated in the training of the ADL classification networks, resulting in models far less susceptible to changes in data distributions. Extensive experiments on the Synthetic-to-Real benchmark Sims4Action demonstrate the potential of the domain generation paradigm for cross-domain ADL recognition, setting new state-of-the-art results. Our code is publicly available at https://github.com/Zrrr1997/syn2real_DG
Video Object Segmentation (VOS) is fundamental to video understanding. Transformer-based methods show significant performance improvement on semi-supervised VOS. However, existing work faces challenges segmenting visually similar objects in close proximity of each other. In this paper, we propose a novel Bilateral Attention Transformer in Motion-Appearance Neighboring space (BATMAN) for semi-supervised VOS. It captures object motion in the video via a novel optical flow calibration module that fuses the segmentation mask with optical flow estimation to improve within-object optical flow smoothness and reduce noise at object boundaries. This calibrated optical flow is then employed in our novel bilateral attention, which computes the correspondence between the query and reference frames in the neighboring bilateral space considering both motion and appearance. Extensive experiments validate the effectiveness of BATMAN architecture by outperforming all existing state-of-the-art on all four popular VOS benchmarks: Youtube-VOS 2019 (85.0%), Youtube-VOS 2018 (85.3%), DAVIS 2017Val/Testdev (86.2%/82.2%), and DAVIS 2016 (92.5%).
Movement synchrony refers to the dynamic temporal connection between the motions of interacting people. The applications of movement synchrony are wide and broad. For example, as a measure of coordination between teammates, synchrony scores are often reported in sports. The autism community also identifies movement synchrony as a key indicator of children's social and developmental achievements. In general, raw video recordings are often used for movement synchrony estimation, with the drawback that they may reveal people's identities. Furthermore, such privacy concern also hinders data sharing, one major roadblock to a fair comparison between different approaches in autism research. To address the issue, this paper proposes an ensemble method for movement synchrony estimation, one of the first deep-learning-based methods for automatic movement synchrony assessment under privacy-preserving conditions. Our method relies entirely on publicly shareable, identity-agnostic secondary data, such as skeleton data and optical flow. We validate our method on two datasets: (1) PT13 dataset collected from autism therapy interventions and (2) TASD-2 dataset collected from synchronized diving competitions. In this context, our method outperforms its counterpart approaches, both deep neural networks and alternatives.
Optical flow computation is essential in the early stages of the video processing pipeline. This paper focuses on a less explored problem in this area, the 360$^\circ$ optical flow estimation using deep neural networks to support increasingly popular VR applications. To address the distortions of panoramic representations when applying convolutional neural networks, we propose a novel multi-projection fusion framework that fuses the optical flow predicted by the models trained using different projection methods. It learns to combine the complementary information in the optical flow results under different projections. We also build the first large-scale panoramic optical flow dataset to support the training of neural networks and the evaluation of panoramic optical flow estimation methods. The experimental results on our dataset demonstrate that our method outperforms the existing methods and other alternative deep networks that were developed for processing 360{\deg} content.
Previous methods based on 3DCNN, convLSTM, or optical flow have achieved great success in video salient object detection (VSOD). However, they still suffer from high computational costs or poor quality of the generated saliency maps. To solve these problems, we design a space-time memory (STM)-based network, which extracts useful temporal information of the current frame from adjacent frames as the temporal branch of VSOD. Furthermore, previous methods only considered single-frame prediction without temporal association. As a result, the model may not focus on the temporal information sufficiently. Thus, we initially introduce object motion prediction between inter-frame into VSOD. Our model follows standard encoder--decoder architecture. In the encoding stage, we generate high-level temporal features by using high-level features from the current and its adjacent frames. This approach is more efficient than the optical flow-based methods. In the decoding stage, we propose an effective fusion strategy for spatial and temporal branches. The semantic information of the high-level features is used to fuse the object details in the low-level features, and then the spatiotemporal features are obtained step by step to reconstruct the saliency maps. Moreover, inspired by the boundary supervision commonly used in image salient object detection (ISOD), we design a motion-aware loss for predicting object boundary motion and simultaneously perform multitask learning for VSOD and object motion prediction, which can further facilitate the model to extract spatiotemporal features accurately and maintain the object integrity. Extensive experiments on several datasets demonstrated the effectiveness of our method and can achieve state-of-the-art metrics on some datasets. The proposed model does not require optical flow or other preprocessing, and can reach a speed of nearly 100 FPS during inference.
Video frame interpolation is a classic and challenging low-level computer vision task. Recently, deep learning based methods have achieved impressive results, and it has been proven that optical flow based methods can synthesize frames with higher quality. However, most flow-based methods assume a line trajectory with a constant velocity between two input frames. Only a little work enforces predictions with curvilinear trajectory, but this requires more than two frames as input to estimate the acceleration, which takes more time and memory to execute. To address this problem, we propose an arc trajectory based model (ATCA), which learns motion prior from only two consecutive frames and also is lightweight. Experiments show that our approach performs better than many SOTA methods with fewer parameters and faster inference speed.
Temporal consistency is the key challenge of video depth estimation. Previous works are based on additional optical flow or camera poses, which is time-consuming. By contrast, we derive consistency with less information. Since videos inherently exist with heavy temporal redundancy, a missing frame could be recovered from neighboring ones. Inspired by this, we propose the frame masking network (FMNet), a spatial-temporal transformer network predicting the depth of masked frames based on their neighboring frames. By reconstructing masked temporal features, the FMNet can learn intrinsic inter-frame correlations, which leads to consistency. Compared with prior arts, experimental results demonstrate that our approach achieves comparable spatial accuracy and higher temporal consistency without any additional information. Our work provides a new perspective on consistent video depth estimation. Our official project page is https://github.com/RaymondWang987/FMNet.
Deepfakes are the synthesized digital media in order to create ultra-realistic fake videos to trick the spectator. Deep generative algorithms, such as, Generative Adversarial Networks(GAN) are widely used to accomplish such tasks. This approach synthesizes pseudo-realistic contents that are very difficult to distinguish by traditional detection methods. In most cases, Convolutional Neural Network(CNN) based discriminators are being used for detecting such synthesized media. However, it emphasise primarily on the spatial attributes of individual video frames, thereby fail to learn the temporal information from their inter-frame relations. In this paper, we leveraged an optical flow based feature extraction approach to extract the temporal features, which are then fed to a hybrid model for classification. This hybrid model is based on the combination of CNN and recurrent neural network (RNN) architectures. The hybrid model provides effective performance on open source data-sets such as, DFDC, FF++ and Celeb-DF. This proposed method shows an accuracy of 66.26%, 91.21% and 79.49% in DFDC, FF++, and Celeb-DF respectively with a very reduced No of sample size of approx 100 samples(frames). This promises early detection of fake contents compared to existing modalities.
Modern 3D computer vision leverages learning to boost geometric reasoning, mapping image data to classical structures such as cost volumes or epipolar constraints to improve matching. These architectures are specialized according to the particular problem, and thus require significant task-specific tuning, often leading to poor domain generalization performance. Recently, generalist Transformer architectures have achieved impressive results in tasks such as optical flow and depth estimation by encoding geometric priors as inputs rather than as enforced constraints. In this paper, we extend this idea and propose to learn an implicit, multi-view consistent scene representation, introducing a series of 3D data augmentation techniques as a geometric inductive prior to increase view diversity. We also show that introducing view synthesis as an auxiliary task further improves depth estimation. Our Depth Field Networks (DeFiNe) achieve state-of-the-art results in stereo and video depth estimation without explicit geometric constraints, and improve on zero-shot domain generalization by a wide margin.
Existing video frame interpolation methods can only interpolate the frame at a given intermediate time-step, e.g. 1/2. In this paper, we aim to explore a more generalized kind of video frame interpolation, that at an arbitrary time-step. To this end, we consider processing different time-steps with adaptively generated convolutional kernels in a unified way with the help of meta-learning. Specifically, we develop a dual meta-learned frame interpolation framework to synthesize intermediate frames with the guidance of context information and optical flow as well as taking the time-step as side information. First, a content-aware meta-learned flow refinement module is built to improve the accuracy of the optical flow estimation based on the down-sampled version of the input frames. Second, with the refined optical flow and the time-step as the input, a motion-aware meta-learned frame interpolation module generates the convolutional kernels for every pixel used in the convolution operations on the feature map of the coarse warped version of the input frames to generate the predicted frame. Extensive qualitative and quantitative evaluations, as well as ablation studies, demonstrate that, via introducing meta-learning in our framework in such a well-designed way, our method not only achieves superior performance to state-of-the-art frame interpolation approaches but also owns an extended capacity to support the interpolation at an arbitrary time-step.
Video deblurring is a highly under-constrained problem due to the spatially and temporally varying blur. An intuitive approach for video deblurring includes two steps: a) detecting the blurry region in the current frame; b) utilizing the information from clear regions in adjacent frames for current frame deblurring. To realize this process, our idea is to detect the pixel-wise blur level of each frame and combine it with video deblurring. To this end, we propose a novel framework that utilizes the motion magnitude prior (MMP) as guidance for efficient deep video deblurring. Specifically, as the pixel movement along its trajectory during the exposure time is positively correlated to the level of motion blur, we first use the average magnitude of optical flow from the high-frequency sharp frames to generate the synthetic blurry frames and their corresponding pixel-wise motion magnitude maps. We then build a dataset including the blurry frame and MMP pairs. The MMP is then learned by a compact CNN by regression. The MMP consists of both spatial and temporal blur level information, which can be further integrated into an efficient recurrent neural network (RNN) for video deblurring. We conduct intensive experiments to validate the effectiveness of the proposed methods on the public datasets.
Semantic segmentation for autonomous driving should be robust against various in-the-wild environments. Nighttime semantic segmentation is especially challenging due to a lack of annotated nighttime images and a large domain gap from daytime images with sufficient annotation. In this paper, we propose a novel GPS-based training framework for nighttime semantic segmentation. Given GPS-aligned pairs of daytime and nighttime images, we perform cross-domain correspondence matching to obtain pixel-level pseudo supervision. Moreover, we conduct flow estimation between daytime video frames and apply GPS-based scaling to acquire another pixel-level pseudo supervision. Using these pseudo supervisions with a confidence map, we train a nighttime semantic segmentation network without any annotation from nighttime images. Experimental results demonstrate the effectiveness of the proposed method on several nighttime semantic segmentation datasets. Our source code is available at https://github.com/jimmy9704/GPS-GLASS.
Snapshot compressive imaging (SCI) encodes high-speed scene video into a snapshot measurement and then computationally makes reconstructions, allowing for efficient high-dimensional data acquisition. Numerous algorithms, ranging from regularization-based optimization and deep learning, are being investigated to improve reconstruction quality, but they are still limited by the ill-posed and information-deficient nature of the standard SCI paradigm. To overcome these drawbacks, we propose a new key frames assisted hybrid encoding paradigm for compressive video sensing, termed KH-CVS, that alternatively captures short-exposure key frames without coding and long-exposure encoded compressive frames to jointly reconstruct photorealistic video. With the use of optical flow and spatial warping, a deep convolutional neural network framework is constructed to integrate the benefits of these two types of frames. Extensive experiments on both simulations and real data from the prototype we developed verify the superiority of the proposed method.