Recent works have shown that optical flow can be learned by deep networks from unlabelled image pairs based on brightness constancy assumption and smoothness prior. Current approaches additionally impose an augmentation regularization term for continual self-supervision, which has been proved to be effective on difficult matching regions. However, this method also amplify the inevitable mismatch in unsupervised setting, blocking the learning process towards optimal solution. To break the dilemma, we propose a novel mutual distillation framework to transfer reliable knowledge back and forth between the teacher and student networks for alternate improvement. Concretely, taking estimation of off-the-shelf unsupervised approach as pseudo labels, our insight locates at defining a confidence selection mechanism to extract relative good matches, and then add diverse data augmentation for distilling adequate and reliable knowledge from teacher to student. Thanks to the decouple nature of our method, we can choose a stronger student architecture for sufficient learning. Finally, better student prediction is adopted to transfer knowledge back to the efficient teacher without additional costs in real deployment. Rather than formulating it as a supervised task, we find that introducing an extra unsupervised term for multi-target learning achieves best final results. Extensive experiments show that our approach, termed MDFlow, achieves state-of-the-art real-time accuracy and generalization ability on challenging benchmarks. Code is available at https://github.com/ltkong218/MDFlow.
Due to the problem of performance constraints of unsupervised video object detection, its large-scale application is limited. In response to this pain point, we propose another excellent method to solve this problematic point. By incorporating motion characterization in unsupervised video object detection, detection accuracy is improved while reducing the computational amount of the network. The whole network structure consists of dual-stream network, motion guidance module, and multi-scale progressive fusion module. The appearance and motion representations of the detection target are obtained through a dual-stream network. Then, the semantic features of the motion representation are obtained through the local attention mechanism in the motion guidance module to obtain the high-level semantic features of the appearance representation. The multi-scale progressive fusion module then fuses the features of different deep semantic features in the dual-stream network further to improve the detection effect of the overall network. We have conducted numerous experiments on the three datasets of DAVIS 16, FBMS, and ViSal. The verification results show that the proposed method achieves superior accuracy and performance and proves the superiority and robustness of the algorithm.
Temporal action localization (TAL) aims to detect the boundary and identify the class of each action instance in a long untrimmed video. Current approaches treat video frames homogeneously, and tend to give background and key objects excessive attention. This limits their sensitivity to localize action boundaries. To this end, we propose a prior-enhanced temporal action localization method (PETAL), which only takes in RGB input and incorporates action subjects as priors. This proposal leverages action subjects' information with a plug-and-play subject-aware spatial attention module (SA-SAM) to generate an aggregated and subject-prioritized representation. Experimental results on THUMOS-14 and ActivityNet-1.3 datasets demonstrate that the proposed PETAL achieves competitive performance using only RGB features, e.g., boosting mAP by 2.41% or 0.25% over the state-of-the-art approach that uses RGB features or with additional optical flow features on the THUMOS-14 dataset.
We present a unified formulation and model for three motion and 3D perception tasks: optical flow, rectified stereo matching and unrectified stereo depth estimation from posed images. Unlike previous specialized architectures for each specific task, we formulate all three tasks as a unified dense correspondence matching problem, which can be solved with a single model by directly comparing feature similarities. Such a formulation calls for discriminative feature representations, which we achieve using a Transformer, in particular the cross-attention mechanism. We demonstrate that cross-attention enables integration of knowledge from another image via cross-view interactions, which greatly improves the quality of the extracted features. Our unified model naturally enables cross-task transfer since the model architecture and parameters are shared across tasks. We outperform RAFT with our unified model on the challenging Sintel dataset, and our final model that uses a few additional task-specific refinement steps outperforms or compares favorably to recent state-of-the-art methods on 10 popular flow, stereo and depth datasets, while being simpler and more efficient in terms of model design and inference speed.
Existing correspondence datasets for two-dimensional (2D) cartoon suffer from simple frame composition and monotonic movements, making them insufficient to simulate real animations. In this work, we present a new 2D animation visual correspondence dataset, AnimeRun, by converting open source three-dimensional (3D) movies to full scenes in 2D style, including simultaneous moving background and interactions of multiple subjects. Our analyses show that the proposed dataset not only resembles real anime more in image composition, but also possesses richer and more complex motion patterns compared to existing datasets. With this dataset, we establish a comprehensive benchmark by evaluating several existing optical flow and segment matching methods, and analyze shortcomings of these methods on animation data. Data, code and other supplementary materials are available at https://lisiyao21.github.io/projects/AnimeRun.
Drone-camera based human activity recognition (HAR) has received significant attention from the computer vision research community in the past few years. A robust and efficient HAR system has a pivotal role in fields like video surveillance, crowd behavior analysis, sports analysis, and human-computer interaction. What makes it challenging are the complex poses, understanding different viewpoints, and the environmental scenarios where the action is taking place. To address such complexities, in this paper, we propose a novel Sparse Weighted Temporal Fusion (SWTF) module to utilize sparsely sampled video frames for obtaining global weighted temporal fusion outcome. The proposed SWTF is divided into two components. First, a temporal segment network that sparsely samples a given set of frames. Second, weighted temporal fusion, that incorporates a fusion of feature maps derived from optical flow, with raw RGB images. This is followed by base-network, which comprises a convolutional neural network module along with fully connected layers that provide us with activity recognition. The SWTF network can be used as a plug-in module to the existing deep CNN architectures, for optimizing them to learn temporal information by eliminating the need for a separate temporal stream. It has been evaluated on three publicly available benchmark datasets, namely Okutama, MOD20, and Drone-Action. The proposed model has received an accuracy of 72.76%, 92.56%, and 78.86% on the respective datasets thereby surpassing the previous state-of-the-art performances by a significant margin.
The increasing number of surveillance cameras and security concerns have made automatic violent activity detection from surveillance footage an active area for research. Modern deep learning methods have achieved good accuracy in violence detection and proved to be successful because of their applicability in intelligent surveillance systems. However, the models are computationally expensive and large in size because of their inefficient methods for feature extraction. This work presents a novel architecture for violence detection called Two-stream Multi-dimensional Convolutional Network (2s-MDCN), which uses RGB frames and optical flow to detect violence. Our proposed method extracts temporal and spatial information independently by 1D, 2D, and 3D convolutions. Despite combining multi-dimensional convolutional networks, our models are lightweight and efficient due to reduced channel capacity, yet they learn to extract meaningful spatial and temporal information. Additionally, combining RGB frames and optical flow yields 2.2% more accuracy than a single RGB stream. Regardless of having less complexity, our models obtained state-of-the-art accuracy of 89.7% on the largest violence detection benchmark dataset.
Generic motion understanding from video involves not only tracking objects, but also perceiving how their surfaces deform and move. This information is useful to make inferences about 3D shape, physical properties and object interactions. While the problem of tracking arbitrary physical points on surfaces over longer video clips has received some attention, no dataset or benchmark for evaluation existed, until now. In this paper, we first formalize the problem, naming it tracking any point (TAP). We introduce a companion benchmark, TAP-Vid, which is composed of both real-world videos with accurate human annotations of point tracks, and synthetic videos with perfect ground-truth point tracks. Central to the construction of our benchmark is a novel semi-automatic crowdsourced pipeline which uses optical flow estimates to compensate for easier, short-term motion like camera shake, allowing annotators to focus on harder sections of video. We validate our pipeline on synthetic data and propose a simple end-to-end point tracking model TAP-Net, showing that it outperforms all prior methods on our benchmark when trained on synthetic data.
Flow-guided synthesis provides a common framework for frame interpolation, where optical flow is estimated to guide the synthesis of intermediate frames between consecutive inputs. In this paper, we present UPR-Net, a novel Unified Pyramid Recurrent Network for frame interpolation. Cast in a flexible pyramid framework, UPR-Net exploits lightweight recurrent modules for both bi-directional flow estimation and intermediate frame synthesis. At each pyramid level, it leverages estimated bi-directional flow to generate forward-warped representations for frame synthesis; across pyramid levels, it enables iterative refinement for both optical flow and intermediate frame. In particular, we show that our iterative synthesis strategy can significantly improve the robustness of frame interpolation on large motion cases. Despite being extremely lightweight (1.7M parameters), our base version of UPR-Net achieves excellent performance on a large range of benchmarks. Code and trained models of our UPR-Net series are available at: https://github.com/srcn-ivl/UPR-Net.
Learning to localize the sound source in videos without explicit annotations is a novel area of audio-visual research. Existing work in this area focuses on creating attention maps to capture the correlation between the two modalities to localize the source of the sound. In a video, oftentimes, the objects exhibiting movement are the ones generating the sound. In this work, we capture this characteristic by modeling the optical flow in a video as a prior to better aid in localizing the sound source. We further demonstrate that the addition of flow-based attention substantially improves visual sound source localization. Finally, we benchmark our method on standard sound source localization datasets and achieve state-of-the-art performance on the Soundnet Flickr and VGG Sound Source datasets. Code: https://github.com/denfed/heartheflow.
Micro-expression generation aims at transfering the expression from the driving videos to the source images, which can be viewed as a motion transfer task. Recently, several works have been proposed to tackle this problem and achieve great performance. However, due to the intrinsic complexity of the face motion and different attributes of face regions, the task still remains challenging. In this paper, we propose an end-to-end unsupervised motion transfer network to tackle this challenge. As the motion of the face is non-rigid, we adopt an effective and flexible thin-plate spline motion estimation method to estimate the optical flow of the face motion. What's more, we find that several faces with eyeglasses show weird deformation in motion transfering. Thus, we introduce face parsing method to pay specific attention to the eyeglasses regions to ensure the reasonability of the deformation. We conduct several experiments on the provided datasets of the ACM MM 2022 micro-expression grand challenge (MEGC2022) and compare our method with several other typical methods. In comparison, our method shows the best performance. We (Team: USTC-IAT-United) also compare our method with other competitors' in MEGC2022, and the expert evaluation results show that our method performs best, which verifies the effectiveness of our method. Our code is available at https://github.com/HowToNameMe/micro-expression
Spiking Neural Networks (SNNs) are bio-plausible models that hold great potential for realizing energy-efficient implementations of sequential tasks on resource-constrained edge devices. However, commercial edge platforms based on standard GPUs are not optimized to deploy SNNs, resulting in high energy and latency. While analog In-Memory Computing (IMC) platforms can serve as energy-efficient inference engines, they are accursed by the immense energy, latency, and area requirements of high-precision ADCs (HP-ADC), overshadowing the benefits of in-memory computations. We propose a hardware/software co-design methodology to deploy SNNs into an ADC-Less IMC architecture using sense-amplifiers as 1-bit ADCs replacing conventional HP-ADCs and alleviating the above issues. Our proposed framework incurs minimal accuracy degradation by performing hardware-aware training and is able to scale beyond simple image classification tasks to more complex sequential regression tasks. Experiments on complex tasks of optical flow estimation and gesture recognition show that progressively increasing the hardware awareness during SNN training allows the model to adapt and learn the errors due to the non-idealities associated with ADC-Less IMC. Also, the proposed ADC-Less IMC offers significant energy and latency improvements, $2-7\times$ and $8.9-24.6\times$, respectively, depending on the SNN model and the workload, compared to HP-ADC IMC.
We provide a new non-invasive, easy-to-scale for large amounts of subjects and a remotely accessible method for (hidden) emotion detection from videos of human faces. Our approach combines face manifold detection for accurate location of the face in the video with local face manifold embedding to create a common domain for the measurements of muscle micro-movements that is invariant to the movement of the subject in the video. In the next step, we employ the Digital Image Speckle Correlation (DISC) and the optical flow algorithm to compute the pattern of micro-movements in the face. The corresponding vector field is mapped back to the original space and superimposed on the original frames of the videos. Hence, the resulting videos include additional information about the direction of the movement of the muscles in the face. We take the publicly available CK++ dataset of visible emotions and add to it videos of the same format but with hidden emotions. We process all the videos using our micro-movement detection and use the results to train a state-of-the-art network for emotions classification from videos -- Frame Attention Network (FAN). Although the original FAN model achieves very high out-of-sample performance on the original CK++ videos, it does not perform so well on hidden emotions videos. The performance improves significantly when the model is trained and tested on videos with the vector fields of muscle movements. Intuitively, the corresponding arrows serve as edges in the image that are easily captured by the convolutions filters in the FAN network.
In this report, we present our optical flow approach, MS-RAFT+, that won the Robust Vision Challenge 2022. It is based on the MS-RAFT method, which successfully integrates several multi-scale concepts into single-scale RAFT. Our approach extends this method by exploiting an additional finer scale for estimating the flow, which is made feasible by on-demand cost computation. This way, it can not only operate at half the original resolution, but also use MS-RAFT's shared convex upsampler to obtain full resolution flow. Moreover, our approach relies on an adjusted fine-tuning scheme during training. This in turn aims at improving the generalization across benchmarks. Among all participating methods in the Robust Vision Challenge, our approach ranks first on VIPER and second on KITTI, Sintel, and Middlebury, resulting in the first place of the overall ranking.
With the continuous research on Deepfake forensics, recent studies have attempted to provide the fine-grained localization of forgeries, in addition to the coarse classification at the video-level. However, the detection and localization performance of existing Deepfake forensic methods still have plenty of room for further improvement. In this work, we propose a Spatial-Temporal Deepfake Detection and Localization (ST-DDL) network that simultaneously explores spatial and temporal features for detecting and localizing forged regions. Specifically, we design a new Anchor-Mesh Motion (AMM) algorithm to extract temporal (motion) features by modeling the precise geometric movements of the facial micro-expression. Compared with traditional motion extraction methods (e.g., optical flow) designed to simulate large-moving objects, our proposed AMM could better capture the small-displacement facial features. The temporal features and the spatial features are then fused in a Fusion Attention (FA) module based on a Transformer architecture for the eventual Deepfake forensic tasks. The superiority of our ST-DDL network is verified by experimental comparisons with several state-of-the-art competitors, in terms of both video- and pixel-level detection and localization performance. Furthermore, to impel the future development of Deepfake forensics, we build a public forgery dataset consisting of 6000 videos, with many new features such as using widely-used commercial software (e.g., After Effects) for the production, providing online social networks transmitted versions, and splicing multi-source videos. The source code and dataset are available at https://github.com/HighwayWu/ST-DDL.
We propose a bootstrapping framework to enhance human optical flow and pose. We show that, for videos involving humans in scenes, we can improve both the optical flow and the pose estimation quality of humans by considering the two tasks at the same time. We enhance optical flow estimates by fine-tuning them to fit the human pose estimates and vice versa. In more detail, we optimize the pose and optical flow networks to, at inference time, agree with each other. We show that this results in state-of-the-art results on the Human 3.6M and 3D Poses in the Wild datasets, as well as a human-related subset of the Sintel dataset, both in terms of pose estimation accuracy and the optical flow accuracy at human joint locations. Code available at https://github.com/ubc-vision/bootstrapping-human-optical-flow-and-pose
Synthetic datasets are often used to pretrain end-to-end optical flow networks, due to the lack of a large amount of labeled, real-scene data. But major drops in accuracy occur when moving from synthetic to real scenes. How do we better transfer the knowledge learned from synthetic to real domains? To this end, we propose CLIP-FLow, a semi-supervised iterative pseudo-labeling framework to transfer the pretraining knowledge to the target real domain. We leverage large-scale, unlabeled real data to facilitate transfer learning with the supervision of iteratively updated pseudo-ground truth labels, bridging the domain gap between the synthetic and the real. In addition, we propose a contrastive flow loss on reference features and the warped features by pseudo ground truth flows, to further boost the accurate matching and dampen the mismatching due to motion, occlusion, or noisy pseudo labels. We adopt RAFT as the backbone and obtain an F1-all error of 4.11%, i.e. a 19% error reduction from RAFT (5.10%) and ranking 2$^{nd}$ place at submission on the KITTI 2015 benchmark. Our framework can also be extended to other models, e.g. CRAFT, reducing the F1-all error from 4.79% to 4.66% on KITTI 2015 benchmark.
The authors present a novel face tracking approach where optical flow information is incorporated into a modified version of the Viola Jones detection algorithm. In the original algorithm, detection is static, as information from previous frames is not considered. In addition, candidate windows have to pass all stages of the classification cascade, otherwise they are discarded as containing no face. In contrast, the proposed tracker preserves information about the number of classification stages passed by each window. Such information is used to build a likelihood map, which represents the probability of having a face located at that position. Tracking capabilities are provided by extrapolating the position of the likelihood map to the next frame by optical flow computation. The proposed algorithm works in real time on a standard laptop. The system is verified on the Boston Head Tracking Database, showing that the proposed algorithm outperforms the standard Viola Jones detector in terms of detection rate and stability of the output bounding box, as well as including the capability to deal with occlusions. The authors also evaluate two recently published face detectors based on convolutional networks and deformable part models with their algorithm showing a comparable accuracy at a fraction of the computation time.
Videos shot by laymen using hand-held cameras contain undesirable shaky motion. Estimating the global motion between successive frames, in a manner not influenced by moving objects, is central to many video stabilization techniques, but poses significant challenges. A large body of work uses 2D affine transformations or homography for the global motion. However, in this work, we introduce a more general representation scheme, which adapts any existing optical flow network to ignore the moving objects and obtain a spatially smooth approximation of the global motion between video frames. We achieve this by a knowledge distillation approach, where we first introduce a low pass filter module into the optical flow network to constrain the predicted optical flow to be spatially smooth. This becomes our student network, named as \textsc{GlobalFlowNet}. Then, using the original optical flow network as the teacher network, we train the student network using a robust loss function. Given a trained \textsc{GlobalFlowNet}, we stabilize videos using a two stage process. In the first stage, we correct the instability in affine parameters using a quadratic programming approach constrained by a user-specified cropping limit to control loss of field of view. In the second stage, we stabilize the video further by smoothing global motion parameters, expressed using a small number of discrete cosine transform coefficients. In extensive experiments on a variety of different videos, our technique outperforms state of the art techniques in terms of subjective quality and different quantitative measures of video stability. The source code is publicly available at \href{https://github.com/GlobalFlowNet/GlobalFlowNet}{https://github.com/GlobalFlowNet/GlobalFlowNet}
Current adversarial attacks for motion estimation (optical flow) optimize small per-pixel perturbations, which are unlikely to appear in the real world. In contrast, we exploit a real-world weather phenomenon for a novel attack with adversarially optimized snow. At the core of our attack is a differentiable renderer that consistently integrates photorealistic snowflakes with realistic motion into the 3D scene. Through optimization we obtain adversarial snow that significantly impacts the optical flow while being indistinguishable from ordinary snow. Surprisingly, the impact of our novel attack is largest on methods that previously showed a high robustness to small L_p perturbations.