There has been tremendous progress in multimodal Large Language Models (LLMs). Recent works have extended these models to video input with promising instruction following capabilities. However, an important missing piece is temporal localization. These models cannot accurately answer the "When?" questions. We identify three key aspects that limit their temporal localization capabilities: (i) time representation, (ii) architecture, and (iii) data. We address these shortcomings by proposing Language Instructed Temporal-Localization Assistant (LITA) with the following features: (1) We introduce time tokens that encode timestamps relative to the video length to better represent time in videos. (2) We introduce SlowFast tokens in the architecture to capture temporal information at fine temporal resolution. (3) We emphasize temporal localization data for LITA. In addition to leveraging existing video datasets with timestamps, we propose a new task, Reasoning Temporal Localization (RTL), along with the dataset, ActivityNet-RTL, for learning and evaluating this task. Reasoning temporal localization requires both the reasoning and temporal localization of Video LLMs. LITA demonstrates strong performance on this challenging task, nearly doubling the temporal mean intersection-over-union (mIoU) of baselines. In addition, we show that our emphasis on temporal localization also substantially improves video-based text generation compared to existing Video LLMs, including a 36% relative improvement of Temporal Understanding. Code is available at: https://github.com/NVlabs/LITA
Understanding human actions from body poses is critical for assistive robots sharing space with humans in order to make informed and safe decisions about the next interaction. However, precise temporal localization and annotation of activity sequences is time-consuming and the resulting labels are often noisy. If not effectively addressed, label noise negatively affects the model's training, resulting in lower recognition quality. Despite its importance, addressing label noise for skeleton-based action recognition has been overlooked so far. In this study, we bridge this gap by implementing a framework that augments well-established skeleton-based human action recognition methods with label-denoising strategies from various research areas to serve as the initial benchmark. Observations reveal that these baselines yield only marginal performance when dealing with sparse skeleton data. Consequently, we introduce a novel methodology, NoiseEraSAR, which integrates global sample selection, co-teaching, and Cross-Modal Mixture-of-Experts (CM-MOE) strategies, aimed at mitigating the adverse impacts of label noise. Our proposed approach demonstrates better performance on the established benchmark, setting new state-of-the-art standards. The source code for this study is accessible at https://github.com/xuyizdby/NoiseEraSAR.
In this work, we focus on label efficient learning for video action detection. We develop a novel semi-supervised active learning approach which utilizes both labeled as well as unlabeled data along with informative sample selection for action detection. Video action detection requires spatio-temporal localization along with classification, which poses several challenges for both active learning informative sample selection as well as semi-supervised learning pseudo label generation. First, we propose NoiseAug, a simple augmentation strategy which effectively selects informative samples for video action detection. Next, we propose fft-attention, a novel technique based on high-pass filtering which enables effective utilization of pseudo label for SSL in video action detection by emphasizing on relevant activity region within a video. We evaluate the proposed approach on three different benchmark datasets, UCF-101-24, JHMDB-21, and Youtube-VOS. First, we demonstrate its effectiveness on video action detection where the proposed approach outperforms prior works in semi-supervised and weakly-supervised learning along with several baseline approaches in both UCF101-24 and JHMDB-21. Next, we also show its effectiveness on Youtube-VOS for video object segmentation demonstrating its generalization capability for other dense prediction tasks in videos. The code and models is publicly available at: \url{https://github.com/AKASH2907/semi-sup-active-learning}.
This work proposes TimeChat, a time-sensitive multimodal large language model specifically designed for long video understanding. Our model incorporates two key architectural contributions: (1) a timestamp-aware frame encoder that binds visual content with the timestamp of each frame, and (2) a sliding video Q-Former that produces a video token sequence of varying lengths to accommodate videos of various durations. Additionally, we construct an instruction-tuning dataset, encompassing 6 tasks and a total of 125K instances, to further enhance TimeChat's instruction-following performance. Experiment results across various video understanding tasks, such as dense captioning, temporal grounding, and highlight detection, demonstrate TimeChat's strong zero-shot temporal localization and reasoning capabilities. For example, it achieves +9.2 F1 score and +2.8 CIDEr on YouCook2, +5.8 HIT@1 on QVHighlights, and +27.5 R@1 (IoU=0.5) on Charades-STA, compared to state-of-the-art video large language models, holding the potential to serve as a versatile video assistant for long-form video comprehension tasks and satisfy realistic user requirements.
Recent studies have shown promising results on utilizing large pre-trained image-language models for video question answering. While these image-language models can efficiently bootstrap the representation learning of video-language models, they typically concatenate uniformly sampled video frames as visual inputs without explicit language-aware, temporal modeling. When only a portion of a video input is relevant to the language query, such uniform frame sampling can often lead to missing important visual cues. Although humans often find a video moment to focus on and rewind the moment to answer questions, training a query-aware video moment localizer often requires expensive annotations and high computational costs. To address this issue, we propose Self-Chained Video Localization-Answering (SeViLA), a novel framework that leverages a single image-language model (BLIP-2) to tackle both temporal keyframe localization and QA on videos. SeViLA framework consists of two modules: Localizer and Answerer, where both are parameter-efficiently fine-tuned from BLIP-2. We propose two ways of chaining these modules for cascaded inference and self-refinement. First, in the forward chain, the Localizer finds multiple language-aware keyframes in a video, which the Answerer uses to predict the answer. Second, in the reverse chain, the Answerer generates keyframe pseudo-labels to refine the Localizer, alleviating the need for expensive video moment localization annotations. Our SeViLA framework outperforms several strong baselines on 5 challenging video QA and event prediction benchmarks, and achieves the state-of-the-art in both fine-tuning (NExT-QA, STAR) and zero-shot (NExT-QA, STAR, How2QA, VLEP) settings. We also analyze the impact of Localizer, comparisons of Localizer with other temporal localization models, pre-training/self-refinement of Localizer, and varying the number of keyframes.
While large-scale image-text pretrained models such as CLIP have been used for multiple video-level tasks on trimmed videos, their use for temporal localization in untrimmed videos is still a relatively unexplored task. We design a new approach for this called UnLoc, which uses pretrained image and text towers, and feeds tokens to a video-text fusion model. The output of the fusion module are then used to construct a feature pyramid in which each level connects to a head to predict a per-frame relevancy score and start/end time displacements. Unlike previous works, our architecture enables Moment Retrieval, Temporal Localization, and Action Segmentation with a single stage model, without the need for action proposals, motion based pretrained features or representation masking. Unlike specialized models, we achieve state of the art results on all three different localization tasks with a unified approach. Code will be available at: \url{https://github.com/google-research/scenic}.
We evaluate the video understanding capabilities of existing foundation models (FMs) using a carefully designed experiment protocol consisting of three hallmark tasks (action recognition,temporal localization, and spatiotemporal localization), eight datasets well received by the community, and four adaptation methods tailoring an FM for downstream tasks. Furthermore, we jointly profile FMs' efficacy and efficiency when adapting to general video understanding tasks using cost measurements during both training and inference. Our main findings areas follows. First, task-specialized models significantly outperform the seven FMs studied in this work, in sharp contrast to what FMs have achieved in natural language and image understanding. Second, video-native FMs, whose pretraining data mainly contains the video modality, are generally better than image-native FMs in classifying motion-rich videos, localizing actions in time, and understanding a video of more than one action. Third, the video-native FMs can perform well on video tasks under light adaptations to downstream tasks (e.g., freezing the FM backbones), while image-native FMs win in full end-to-end finetuning. The first two observations reveal the need and tremendous opportunities to conduct research on video-focused FMs, and the last confirms that both tasks and adaptation methods matter when it comes to the evaluation of FMs. Our code is released under: https://github.com/tensorflow/models/tree/master/official/projects/videoglue.
We propose a new task and model for dense video object captioning -- detecting, tracking and captioning trajectories of objects in a video. This task unifies spatial and temporal localization in video, whilst also requiring fine-grained visual understanding that is best described by natural language. We propose a unified model, and demonstrate how our end-to-end approach is more accurate and temporally coherent than a multi-stage pipeline combining state-of-the-art detection, tracking, and captioning models. Moreover, we propose a training strategy based on a mixture of disjoint tasks, which allows us to leverage diverse, large-scale datasets which supervise different parts of our model. Although each pretraining task only provides weak supervision, they are complementary and, when combined, result in noteworthy zero-shot ability and serve as strong initialization for additional finetuning to further improve accuracy. We carefully design new metrics capturing all components of our task, and show how we can repurpose existing video grounding datasets (e.g. VidSTG and VLN) for our new task. We show that our model improves upon a number of strong baselines for this new task. Furthermore, we can apply our model to the task of spatial grounding, outperforming prior state-of-the-art on VidSTG and VLN, without explicitly training for it. Code is available at https://github.com/google-research/scenic/tree/main/scenic/projects/densevoc.
Open audio databases such as Xeno-Canto are widely used to build datasets to explore bird song repertoire or to train models for automatic bird sound classification by deep learning algorithms. However, such databases suffer from the fact that bird sounds are weakly labelled: a species name is attributed to each audio recording without timestamps that provide the temporal localization of the bird song of interest. Manual annotations can solve this issue, but they are time consuming, expert-dependent, and cannot run on large datasets. Another solution consists in using a labelling function that automatically segments audio recordings before assigning a label to each segmented audio sample. Although labelling functions were introduced to expedite strong label assignment, their classification performance remains mostly unknown. To address this issue and reduce label noise (wrong label assignment) in large bird song datasets, we introduce a data-centric novel labelling function composed of three successive steps: 1) time-frequency sound unit segmentation, 2) feature computation for each sound unit, and 3) classification of each sound unit as bird song or noise with either an unsupervised DBSCAN algorithm or the supervised BirdNET neural network. The labelling function was optimized, validated, and tested on the songs of 44 West-Palearctic common bird species. We first showed that the segmentation of bird songs alone aggregated from 10% to 83% of label noise depending on the species. We also demonstrated that our labelling function was able to significantly reduce the initial label noise present in the dataset by up to a factor of three. Finally, we discuss different opportunities to design suitable labelling functions to build high-quality animal vocalizations with minimum expert annotation effort.
In this report, we present our approach and empirical results of applying masked autoencoders in two egocentric video understanding tasks, namely, Object State Change Classification and PNR Temporal Localization, of Ego4D Challenge 2022. As team TheSSVL, we ranked 2nd place in both tasks. Our code will be made available.
We aim to learn to temporally localize object state changes and the corresponding state-modifying actions by observing people interacting with objects in long uncurated web videos. We introduce three principal contributions. First, we explore alternative multi-task network architectures and identify a model that enables efficient joint learning of multiple object states and actions such as pouring water and pouring coffee. Second, we design a multi-task self-supervised learning procedure that exploits different types of constraints between objects and state-modifying actions enabling end-to-end training of a model for temporal localization of object states and actions in videos from only noisy video-level supervision. Third, we report results on the large-scale ChangeIt and COIN datasets containing tens of thousands of long (un)curated web videos depicting various interactions such as hole drilling, cream whisking, or paper plane folding. We show that our multi-task model achieves a relative improvement of 40% over the prior single-task methods and significantly outperforms both image-based and video-based zero-shot models for this problem. We also test our method on long egocentric videos of the EPIC-KITCHENS and the Ego4D datasets in a zero-shot setup demonstrating the robustness of our learned model.
Video-Text Pre-training (VTP) aims to learn transferable representations for various downstream tasks from large-scale web videos. To date, almost all existing VTP methods are limited to retrieval-based downstream tasks, e.g., video retrieval, whereas their transfer potentials on localization-based tasks, e.g., temporal grounding, are under-explored. In this paper, we experimentally analyze and demonstrate the incompatibility of current VTP methods with localization tasks, and propose a novel Localization-oriented Video-Text Pre-training framework, dubbed as LocVTP. Specifically, we perform the fine-grained contrastive alignment as a complement to the coarse-grained one by a clip-word correspondence discovery scheme. To further enhance the temporal reasoning ability of the learned feature, we propose a context projection head and a temporal aware contrastive loss to perceive the contextual relationships. Extensive experiments on four downstream tasks across six datasets demonstrate that our LocVTP achieves state-of-the-art performance on both retrieval-based and localization-based tasks. Furthermore, we conduct comprehensive ablation studies and thorough analyses to explore the optimum model designs and training strategies.
Temporal Action Detection(TAD) is a crucial but challenging task in video understanding.It is aimed at detecting both the type and start-end frame for each action instance in a long, untrimmed video.Most current models adopt both RGB and Optical-Flow streams for the TAD task. Thus, original RGB frames must be converted manually into Optical-Flow frames with additional computation and time cost, which is an obstacle to achieve real-time processing. At present, many models adopt two-stage strategies, which would slow the inference speed down and complicatedly tuning on proposals generating.By comparison, we propose a one-stage anchor-free temporal localization method with RGB stream only, in which a novel Newtonian Mechanics-MLP architecture is established. It has comparable accuracy with all existing state-of-the-art models, while surpasses the inference speed of these methods by a large margin. The typical inference speed in this paper is astounding 4.44 video per second on THUMOS14. In applications, because there is no need to convert optical flow, the inference speed will be faster.It also proves that MLP has great potential in downstream tasks such as TAD. The source code is available at https://github.com/BonedDeng/TadML
Video-Language Pretraining (VLP), which aims to learn transferable representation to advance a wide range of video-text downstream tasks, has recently received increasing attention. Best performing works rely on large-scale, 3rd-person video-text datasets, such as HowTo100M. In this work, we exploit the recently released Ego4D dataset to pioneer Egocentric VLP along three directions. (i) We create EgoClip, a 1st-person video-text pretraining dataset comprising 3.8M clip-text pairs well-chosen from Ego4D, covering a large variety of human daily activities. (ii) We propose a novel pretraining objective, dubbed EgoNCE, which adapts video-text contrastive learning to the egocentric domain by mining egocentric-aware positive and negative samples. (iii) We introduce EgoMCQ, a development benchmark that is close to EgoClip and hence can support effective validation and fast exploration of our design decisions in EgoClip and EgoNCE. Furthermore, we demonstrate strong performance on five egocentric downstream tasks across three datasets: video-text retrieval on EPIC-KITCHENS-100; action recognition on Charades-Ego; natural language query, moment query, and object state change classification on Ego4D challenge benchmarks. The dataset and code are available at https://github.com/showlab/EgoVLP.
We present a model for temporally precise action spotting in videos, which uses a dense set of detection anchors, predicting a detection confidence and corresponding fine-grained temporal displacement for each anchor. We experiment with two trunk architectures, both of which are able to incorporate large temporal contexts while preserving the smaller-scale features required for precise localization: a one-dimensional version of a u-net, and a Transformer encoder (TE). We also suggest best practices for training models of this kind, by applying Sharpness-Aware Minimization (SAM) and mixup data augmentation. We achieve a new state-of-the-art on SoccerNet-v2, the largest soccer video dataset of its kind, with marked improvements in temporal localization. Additionally, our ablations show: the importance of predicting the temporal displacements; the trade-offs between the u-net and TE trunks; and the benefits of training with SAM and mixup.
We consider the problem of localizing a spatio-temporal tube in a video corresponding to a given text query. This is a challenging task that requires the joint and efficient modeling of temporal, spatial and multi-modal interactions. To address this task, we propose TubeDETR, a transformer-based architecture inspired by the recent success of such models for text-conditioned object detection. Our model notably includes: (i) an efficient video and text encoder that models spatial multi-modal interactions over sparsely sampled frames and (ii) a space-time decoder that jointly performs spatio-temporal localization. We demonstrate the advantage of our proposed components through an extensive ablation study. We also evaluate our full approach on the spatio-temporal video grounding task and demonstrate improvements over the state of the art on the challenging VidSTG and HC-STVG benchmarks. Code and trained models are publicly available at https://antoyang.github.io/tubedetr.html.
Unsupervised video representation learning has made remarkable achievements in recent years. However, most existing methods are designed and optimized for video classification. These pre-trained models can be sub-optimal for temporal localization tasks due to the inherent discrepancy between video-level classification and clip-level localization. To bridge this gap, we make the first attempt to propose a self-supervised pretext task, coined as Pseudo Action Localization (PAL) to Unsupervisedly Pre-train feature encoders for Temporal Action Localization tasks (UP-TAL). Specifically, we first randomly select temporal regions, each of which contains multiple clips, from one video as pseudo actions and then paste them onto different temporal positions of the other two videos. The pretext task is to align the features of pasted pseudo action regions from two synthetic videos and maximize the agreement between them. Compared to the existing unsupervised video representation learning approaches, our PAL adapts better to downstream TAL tasks by introducing a temporal equivariant contrastive learning paradigm in a temporally dense and scale-aware manner. Extensive experiments show that PAL can utilize large-scale unlabeled video data to significantly boost the performance of existing TAL methods. Our codes and models will be made publicly available at https://github.com/zhang-can/UP-TAL.
Temporal Action Localization (TAL) has experienced remarkable success under the supervised learning paradigm. However, existing TAL methods are rooted in the closed set assumption, which cannot handle the inevitable unknown actions in open-world scenarios. In this paper, we, for the first time, step toward the Open Set TAL (OSTAL) problem and propose a general framework OpenTAL based on Evidential Deep Learning (EDL). Specifically, the OpenTAL consists of uncertainty-aware action classification, actionness prediction, and temporal location regression. With the proposed importance-balanced EDL method, classification uncertainty is learned by collecting categorical evidence majorly from important samples. To distinguish the unknown actions from background video frames, the actionness is learned by the positive-unlabeled learning. The classification uncertainty is further calibrated by leveraging the guidance from the temporal localization quality. The OpenTAL is general to enable existing TAL models for open set scenarios, and experimental results on THUMOS14 and ActivityNet1.3 benchmarks show the effectiveness of our method. The code and pre-trained models are released at https://www.rit.edu/actionlab/opental.
In this work, we focus on semi-supervised learning for video action detection which utilizes both labeled as well as unlabeled data. We propose a simple end-to-end consistency based approach which effectively utilizes the unlabeled data. Video action detection requires both, action class prediction as well as a spatio-temporal localization of actions. Therefore, we investigate two types of constraints, classification consistency, and spatio-temporal consistency. The presence of predominant background and static regions in a video makes it challenging to utilize spatio-temporal consistency for action detection. To address this, we propose two novel regularization constraints for spatio-temporal consistency; 1) temporal coherency, and 2) gradient smoothness. Both these aspects exploit the temporal continuity of action in videos and are found to be effective for utilizing unlabeled videos for action detection. We demonstrate the effectiveness of the proposed approach on two different action detection benchmark datasets, UCF101-24 and JHMDB-21. In addition, we also show the effectiveness of the proposed approach for video object segmentation on the Youtube-VOS which demonstrates its generalization capability The proposed approach achieves competitive performance by using merely 20% of annotations on UCF101-24 when compared with recent fully supervised methods. On UCF101-24, it improves the score by +8.9% and +11% at 0.5 f-mAP and v-mAP respectively, compared to supervised approach.
We consider the task of temporal human action localization in lifestyle vlogs. We introduce a novel dataset consisting of manual annotations of temporal localization for 13,000 narrated actions in 1,200 video clips. We present an extensive analysis of this data, which allows us to better understand how the language and visual modalities interact throughout the videos. We propose a simple yet effective method to localize the narrated actions based on their expected duration. Through several experiments and analyses, we show that our method brings complementary information with respect to previous methods, and leads to improvements over previous work for the task of temporal action localization.