Real-world user-generated videos, especially on platforms like TikTok, often feature rich and intertwined audio visual content. However, existing video captioning benchmarks and models remain predominantly visual centric, overlooking the crucial role of audio in conveying scene dynamics, speaker intent, and narrative context. This lack of omni datasets and lightweight, capable models hampers progress in fine grained, multimodal video understanding. To address these challenges, we introduce UGC-VideoCap, a new benchmark and model framework specifically designed for detailed omnimodal captioning of short form user-generated videos. Unlike prior datasets, UGC-VideoCap emphasizes balanced integration of audio and visual modalities, featuring 1000 TikTok videos annotated through a structured three stage human-in-the-loop pipeline covering audio only, visual only, and joint audio visual semantics. The benchmark also includes 4000 carefully crafted QA pairs probing both unimodal and cross modal understanding. Alongside the dataset, we propose UGC-VideoCaptioner(3B), a 3B parameter captioning model distilled from Gemini 2.5 Flash. Using a novel two-stage training strategy supervised fine tuning followed by Group Relative Policy Optimization (GRPO), our approach enables efficient adaptation from limited data while maintaining competitive performance. Together, our benchmark and model offer a high-quality foundation and a data-efficient solution for advancing omnimodal video captioning in unconstrained real-world UGC settings.
Videos contain a wealth of information, and generating detailed and accurate descriptions in natural language is a key aspect of video understanding. In this paper, we present video-SALMONN 2, an advanced audio-visual large language model (LLM) with low-rank adaptation (LoRA) designed for enhanced video (with paired audio) captioning through directed preference optimisation (DPO). We propose new metrics to evaluate the completeness and accuracy of video descriptions, which are optimised using DPO. To further improve training, we propose a novel multi-round DPO (MrDPO) approach, which involves periodically updating the DPO reference model, merging and re-initialising the LoRA module as a proxy for parameter updates after each training round (1,000 steps), and incorporating guidance from ground-truth video captions to stabilise the process. Experimental results show that MrDPO significantly enhances video-SALMONN 2's captioning accuracy, reducing the captioning error rates by 28\%. The final video-SALMONN 2 model, with just 7 billion parameters, surpasses leading models such as GPT-4o and Gemini-1.5-Pro in video captioning tasks, while maintaining highly competitive performance to the state-of-the-art on widely used video question-answering benchmarks among models of similar size. Codes are available at \href{https://github.com/bytedance/video-SALMONN-2}{https://github.com/bytedance/video-SALMONN-2}.
The training of controllable text-to-video (T2V) models relies heavily on the alignment between videos and captions, yet little existing research connects video caption evaluation with T2V generation assessment. This paper introduces VidCapBench, a video caption evaluation scheme specifically designed for T2V generation, agnostic to any particular caption format. VidCapBench employs a data annotation pipeline, combining expert model labeling and human refinement, to associate each collected video with key information spanning video aesthetics, content, motion, and physical laws. VidCapBench then partitions these key information attributes into automatically assessable and manually assessable subsets, catering to both the rapid evaluation needs of agile development and the accuracy requirements of thorough validation. By evaluating numerous state-of-the-art captioning models, we demonstrate the superior stability and comprehensiveness of VidCapBench compared to existing video captioning evaluation approaches. Verification with off-the-shelf T2V models reveals a significant positive correlation between scores on VidCapBench and the T2V quality evaluation metrics, indicating that VidCapBench can provide valuable guidance for training T2V models. The project is available at https://github.com/VidCapBench/VidCapBench.
Large language and multimodal models (LLMs and LMMs) exhibit strong inference capabilities but are often limited by slow decoding speeds. This challenge is especially acute in LMMs, where visual inputs typically comprise more tokens with lower information density than text -- an issue exacerbated by recent trends toward finer-grained visual tokenizations to boost performance. Speculative decoding has been effective in accelerating LLM inference by using a smaller draft model to generate candidate tokens, which are then selectively verified by the target model, improving speed without sacrificing output quality. While this strategy has been extended to LMMs, existing methods largely overlook the unique properties of visual inputs and depend solely on text-based draft models. In this work, we propose \textbf{FLASH} (Fast Latent-Aware Semi-Autoregressive Heuristics), a speculative decoding framework designed specifically for LMMs, which leverages two key properties of multimodal data to design the draft model. First, to address redundancy in visual tokens, we propose a lightweight latent-aware token compression mechanism. Second, recognizing that visual objects often co-occur within a scene, we employ a semi-autoregressive decoding strategy to generate multiple tokens per forward pass. These innovations accelerate draft decoding while maintaining high acceptance rates, resulting in faster overall inference. Experiments show that FLASH significantly outperforms prior speculative decoding approaches in both unimodal and multimodal settings, achieving up to \textbf{2.68$\times$} speed-up on video captioning and \textbf{2.55$\times$} on visual instruction tuning tasks compared to the original LMM. Our code is available \href{https://github.com/ZihuaEvan/FlashSD/}{[here]}.
Movie Audio Description (AD) aims to narrate visual content during dialogue-free segments, particularly benefiting blind and visually impaired (BVI) audiences. Compared with general video captioning, AD demands plot-relevant narration with explicit character name references, posing unique challenges in movie understanding.To identify active main characters and focus on storyline-relevant regions, we propose FocusedAD, a novel framework that delivers character-centric movie audio descriptions. It includes: (i) a Character Perception Module(CPM) for tracking character regions and linking them to names; (ii) a Dynamic Prior Module(DPM) that injects contextual cues from prior ADs and subtitles via learnable soft prompts; and (iii) a Focused Caption Module(FCM) that generates narrations enriched with plot-relevant details and named characters. To overcome limitations in character identification, we also introduce an automated pipeline for building character query banks. FocusedAD achieves state-of-the-art performance on multiple benchmarks, including strong zero-shot results on MAD-eval-Named and our newly proposed Cinepile-AD dataset. Code and data will be released at https://github.com/Thorin215/FocusedAD .
We present CAT-V (Caption AnyThing in Video), a training-free framework for fine-grained object-centric video captioning that enables detailed descriptions of user-selected objects through time. CAT-V integrates three key components: a Segmenter based on SAMURAI for precise object segmentation across frames, a Temporal Analyzer powered by TRACE-Uni for accurate event boundary detection and temporal analysis, and a Captioner using InternVL-2.5 for generating detailed object-centric descriptions. Through spatiotemporal visual prompts and chain-of-thought reasoning, our framework generates detailed, temporally-aware descriptions of objects' attributes, actions, statuses, interactions, and environmental contexts without requiring additional training data. CAT-V supports flexible user interactions through various visual prompts (points, bounding boxes, and irregular regions) and maintains temporal sensitivity by tracking object states and interactions across different time segments. Our approach addresses limitations of existing video captioning methods, which either produce overly abstract descriptions or lack object-level precision, enabling fine-grained, object-specific descriptions while maintaining temporal coherence and spatial accuracy. The GitHub repository for this project is available at https://github.com/yunlong10/CAT-V
Developing video captioning models is computationally expensive. The dynamic nature of video also complicates the design of multimodal models that can effectively caption these sequences. However, we find that by using minimal computational resources and without complex modifications to address video dynamics, an image-based model can be repurposed to outperform several specialised video captioning systems. Our adapted model demonstrates top tier performance on major benchmarks, ranking 2nd on MSRVTT and MSVD, and 3rd on VATEX. We transform it into a competitive video captioner by post training a typical image captioning model BLIP2 with only 6,000 video text pairs and simply concatenating frames (significantly fewer data than other methods), which use 2.5 to 144 million pairs. From a resource optimization perspective, this video captioning study focuses on three fundamental factors: optimizing model scale, maximizing data efficiency, and incorporating reinforcement learning. This extensive study demonstrates that a lightweight, image based adaptation strategy can rival state-of-the-art video captioning systems, offering a practical solution for low-resource scenarios.
Multimodal foundation models (MFMs) have demonstrated significant success in tasks such as visual captioning, question answering, and image-text retrieval. However, these models face inherent limitations due to their finite internal capacity, which restricts their ability to process extended temporal sequences, a crucial requirement for comprehensive video and audio analysis. To overcome these challenges, we introduce a specialized cognitive module, temporal working memory (TWM), which aims to enhance the temporal modeling capabilities of MFMs. It selectively retains task-relevant information across temporal dimensions, ensuring that critical details are preserved throughout the processing of video and audio content. The TWM uses a query-guided attention approach to focus on the most informative multimodal segments within temporal sequences. By retaining only the most relevant content, TWM optimizes the use of the model's limited capacity, enhancing its temporal modeling ability. This plug-and-play module can be easily integrated into existing MFMs. With our TWM, nine state-of-the-art models exhibit significant performance improvements across tasks such as video captioning, question answering, and video-text retrieval. By enhancing temporal modeling, TWM extends the capability of MFMs to handle complex, time-sensitive data effectively. Our code is available at https://github.com/xid32/NAACL_2025_TWM.
Despite the advancements of Video Large Language Models (VideoLLMs) in various tasks, they struggle with fine-grained temporal understanding, such as Dense Video Captioning (DVC). DVC is a complicated task of describing all events within a video while also temporally localizing them, which integrates multiple fine-grained tasks, including video segmentation, video captioning, and temporal video grounding. Previous VideoLLMs attempt to solve DVC in a single step, failing to utilize their reasoning capability. Moreover, previous training objectives for VideoLLMs do not fully reflect the evaluation metrics, therefore not providing supervision directly aligned to target tasks. To address such a problem, we propose a novel framework named VidChain comprised of Chain-of-Tasks (CoTasks) and Metric-based Direct Preference Optimization (M-DPO). CoTasks decompose a complex task into a sequence of sub-tasks, allowing VideoLLMs to leverage their reasoning capabilities more effectively. M-DPO aligns a VideoLLM with evaluation metrics, providing fine-grained supervision to each task that is well-aligned with metrics. Applied to two different VideoLLMs, VidChain consistently improves their fine-grained video understanding, thereby outperforming previous VideoLLMs on two different DVC benchmarks and also on the temporal video grounding task. Code is available at \url{https://github.com/mlvlab/VidChain}.
Multimodal representation learning, with contrastive learning, plays an important role in the artificial intelligence domain. As an important subfield, video-language representation learning focuses on learning representations using global semantic interactions between pre-defined video-text pairs. However, to enhance and refine such coarse-grained global interactions, more detailed interactions are necessary for fine-grained multimodal learning. In this study, we introduce a new approach that models video-text as game players using multivariate cooperative game theory to handle uncertainty during fine-grained semantic interactions with diverse granularity, flexible combination, and vague intensity. Specifically, we design the Hierarchical Banzhaf Interaction to simulate the fine-grained correspondence between video clips and textual words from hierarchical perspectives. Furthermore, to mitigate the bias in calculations within Banzhaf Interaction, we propose reconstructing the representation through a fusion of single-modal and cross-modal components. This reconstructed representation ensures fine granularity comparable to that of the single-modal representation, while also preserving the adaptive encoding characteristics of cross-modal representation. Additionally, we extend our original structure into a flexible encoder-decoder framework, enabling the model to adapt to various downstream tasks. Extensive experiments on commonly used text-video retrieval, video-question answering, and video captioning benchmarks, with superior performance, validate the effectiveness and generalization of our method.
Evaluation metric of visual captioning is important yet not thoroughly explored. Traditional metrics like BLEU, METEOR, CIDEr, and ROUGE often miss semantic depth, while trained metrics such as CLIP-Score, PAC-S, and Polos are limited in zero-shot scenarios. Advanced Language Model-based metrics also struggle with aligning to nuanced human preferences. To address these issues, we introduce G-VEval, a novel metric inspired by G-Eval and powered by the new GPT-4o. G-VEval uses chain-of-thought reasoning in large multimodal models and supports three modes: reference-free, reference-only, and combined, accommodating both video and image inputs. We also propose MSVD-Eval, a new dataset for video captioning evaluation, to establish a more transparent and consistent framework for both human experts and evaluation metrics. It is designed to address the lack of clear criteria in existing datasets by introducing distinct dimensions of Accuracy, Completeness, Conciseness, and Relevance (ACCR). Extensive results show that G-VEval outperforms existing methods in correlation with human annotations, as measured by Kendall tau-b and Kendall tau-c. This provides a flexible solution for diverse captioning tasks and suggests a straightforward yet effective approach for large language models to understand video content, paving the way for advancements in automated captioning. Codes are available at https://github.com/ztangaj/gveval
Weakly-Supervised Dense Video Captioning (WSDVC) aims to localize and describe all events of interest in a video without requiring annotations of event boundaries. This setting poses a great challenge in accurately locating the temporal location of event, as the relevant supervision is unavailable. Existing methods rely on explicit alignment constraints between event locations and captions, which involve complex event proposal procedures during both training and inference. To tackle this problem, we propose a novel implicit location-caption alignment paradigm by complementary masking, which simplifies the complex event proposal and localization process while maintaining effectiveness. Specifically, our model comprises two components: a dual-mode video captioning module and a mask generation module. The dual-mode video captioning module captures global event information and generates descriptive captions, while the mask generation module generates differentiable positive and negative masks for localizing the events. These masks enable the implicit alignment of event locations and captions by ensuring that captions generated from positively and negatively masked videos are complementary, thereby forming a complete video description. In this way, even under weak supervision, the event location and event caption can be aligned implicitly. Extensive experiments on the public datasets demonstrate that our method outperforms existing weakly-supervised methods and achieves competitive results compared to fully-supervised methods.
Large Language Model (LLM)-based agents have shown promise in procedural tasks, but the potential of multimodal instructions augmented by texts and videos to assist users remains under-explored. To address this gap, we propose the Visually Grounded Text-Video Prompting (VG-TVP) method which is a novel LLM-empowered Multimodal Procedural Planning (MPP) framework. It generates cohesive text and video procedural plans given a specified high-level objective. The main challenges are achieving textual and visual informativeness, temporal coherence, and accuracy in procedural plans. VG-TVP leverages the zero-shot reasoning capability of LLMs, the video-to-text generation ability of the video captioning models, and the text-to-video generation ability of diffusion models. VG-TVP improves the interaction between modalities by proposing a novel Fusion of Captioning (FoC) method and using Text-to-Video Bridge (T2V-B) and Video-to-Text Bridge (V2T-B). They allow LLMs to guide the generation of visually-grounded text plans and textual-grounded video plans. To address the scarcity of datasets suitable for MPP, we have curated a new dataset called Daily-Life Task Procedural Plans (Daily-PP). We conduct comprehensive experiments and benchmarks to evaluate human preferences (regarding textual and visual informativeness, temporal coherence, and plan accuracy). Our VG-TVP method outperforms unimodal baselines on the Daily-PP dataset.
Recent researches on video large language models (VideoLLM) predominantly focus on model architectures and training datasets, leaving the interaction format between the user and the model under-explored. In existing works, users often interact with VideoLLMs by using the entire video and a query as input, after which the model generates a response. This interaction format constrains the application of VideoLLMs in scenarios such as live-streaming comprehension where videos do not end and responses are required in a real-time manner, and also results in unsatisfactory performance on time-sensitive tasks that requires localizing video segments. In this paper, we focus on a video-text duet interaction format. This interaction format is characterized by the continuous playback of the video, and both the user and the model can insert their text messages at any position during the video playback. When a text message ends, the video continues to play, akin to the alternative of two performers in a duet. We construct MMDuetIT, a video-text training dataset designed to adapt VideoLLMs to video-text duet interaction format. We also introduce the Multi-Answer Grounded Video Question Answering (MAGQA) task to benchmark the real-time response ability of VideoLLMs. Trained on MMDuetIT, MMDuet demonstrates that adopting the video-text duet interaction format enables the model to achieve significant improvements in various time-sensitive tasks (76% CIDEr on YouCook2 dense video captioning, 90\% mAP on QVHighlights highlight detection and 25% [email omitted] on Charades-STA temporal video grounding) with minimal training efforts, and also enable VideoLLMs to reply in a real-time manner as the video plays. Code, data and demo are available at: https://github.com/yellow-binary-tree/MMDuet.
Video captioning generate a sentence that describes the video content. Existing methods always require a number of captions (\eg, 10 or 20) per video to train the model, which is quite costly. In this work, we explore the possibility of using only one or very few ground-truth sentences, and introduce a new task named few-supervised video captioning. Specifically, we propose a few-supervised video captioning framework that consists of lexically constrained pseudo-labeling module and keyword-refined captioning module. Unlike the random sampling in natural language processing that may cause invalid modifications (\ie, edit words), the former module guides the model to edit words using some actions (\eg, copy, replace, insert, and delete) by a pretrained token-level classifier, and then fine-tunes candidate sentences by a pretrained language model. Meanwhile, the former employs the repetition penalized sampling to encourage the model to yield concise pseudo-labeled sentences with less repetition, and selects the most relevant sentences upon a pretrained video-text model. Moreover, to keep semantic consistency between pseudo-labeled sentences and video content, we develop the transformer-based keyword refiner with the video-keyword gated fusion strategy to emphasize more on relevant words. Extensive experiments on several benchmarks demonstrate the advantages of the proposed approach in both few-supervised and fully-supervised scenarios. The code implementation is available at https://github.com/mlvccn/PKG_VidCap
The efficacy of video generation models heavily depends on the quality of their training datasets. Most previous video generation models are trained on short video clips, while recently there has been increasing interest in training long video generation models directly on longer videos. However, the lack of such high-quality long videos impedes the advancement of long video generation. To promote research in long video generation, we desire a new dataset with four key features essential for training long video generation models: (1) long videos covering at least 10 seconds, (2) long-take videos without cuts, (3) large motion and diverse contents, and (4) temporally dense captions. To achieve this, we introduce a new pipeline for selecting high-quality long-take videos and generating temporally dense captions. Specifically, we define a set of metrics to quantitatively assess video quality including scene cuts, dynamic degrees, and semantic-level quality, enabling us to filter high-quality long-take videos from a large amount of source videos. Subsequently, we develop a hierarchical video captioning pipeline to annotate long videos with temporally-dense captions. With this pipeline, we curate the first long-take video dataset, LVD-2M, comprising 2 million long-take videos, each covering more than 10 seconds and annotated with temporally dense captions. We further validate the effectiveness of LVD-2M by fine-tuning video generation models to generate long videos with dynamic motions. We believe our work will significantly contribute to future research in long video generation.
Video Large Language Models (Video-LLMs) have demonstrated remarkable capabilities in coarse-grained video understanding, however, they struggle with fine-grained temporal grounding. In this paper, we introduce Grounded-VideoLLM, a novel Video-LLM adept at perceiving and reasoning over specific video moments in a fine-grained manner. We identify that current Video-LLMs have limitations for fine-grained video understanding since they lack effective temporal modeling and timestamp representation. In light of this, we sharpen our model by incorporating (1) an additional temporal stream to encode the relationships between frames and (2) discrete temporal tokens enriched with specific time knowledge to represent timestamps. To optimize the training of Grounded-VideoLLM, we employ a multi-stage training scheme, beginning with simple video-captioning tasks and progressively introducing video temporal grounding tasks of increasing complexity. To further enhance Grounded-VideoLLM's temporal reasoning capability, we also curate a grounded VideoQA dataset by an automatic annotation pipeline. Extensive experiments demonstrate that Grounded-VideoLLM not only excels in fine-grained grounding tasks such as temporal sentence grounding, dense video captioning, and grounded VideoQA, but also shows great potential as a versatile video assistant for general video understanding.
Recent advancements in image captioning have explored text-only training methods to overcome the limitations of paired image-text data. However, existing text-only training methods often overlook the modality gap between using text data during training and employing images during inference. To address this issue, we propose a novel approach called Image-like Retrieval, which aligns text features with visually relevant features to mitigate the modality gap. Our method further enhances the accuracy of generated captions by designing a Fusion Module that integrates retrieved captions with input features. Additionally, we introduce a Frequency-based Entity Filtering technique that significantly improves caption quality. We integrate these methods into a unified framework, which we refer to as IFCap ($\textbf{I}$mage-like Retrieval and $\textbf{F}$requency-based Entity Filtering for Zero-shot $\textbf{Cap}$tioning). Through extensive experimentation, our straightforward yet powerful approach has demonstrated its efficacy, outperforming the state-of-the-art methods by a significant margin in both image captioning and video captioning compared to zero-shot captioning based on text-only training.
The SoccerNet 2024 challenges represent the fourth annual video understanding challenges organized by the SoccerNet team. These challenges aim to advance research across multiple themes in football, including broadcast video understanding, field understanding, and player understanding. This year, the challenges encompass four vision-based tasks. (1) Ball Action Spotting, focusing on precisely localizing when and which soccer actions related to the ball occur, (2) Dense Video Captioning, focusing on describing the broadcast with natural language and anchored timestamps, (3) Multi-View Foul Recognition, a novel task focusing on analyzing multiple viewpoints of a potential foul incident to classify whether a foul occurred and assess its severity, (4) Game State Reconstruction, another novel task focusing on reconstructing the game state from broadcast videos onto a 2D top-view map of the field. Detailed information about the tasks, challenges, and leaderboards can be found at https://www.soccer-net.org, with baselines and development kits available at https://github.com/SoccerNet.
Generating high-quality shooting scripts containing information such as scene and shot language is essential for short drama script generation. We collect 6,660 popular short drama episodes from the Internet, each with an average of 100 short episodes, and the total number of short episodes is about 80,000, with a total duration of about 2,000 hours and totaling 10 terabytes (TB). We perform keyframe extraction and annotation on each episode to obtain about 10,000,000 shooting scripts. We perform 100 script restorations on the extracted shooting scripts based on our self-developed large short drama generation model SkyReels. This leads to a dataset containing 1,000,000,000 pairs of scripts and shooting scripts for short dramas, called SkyScript-100M. We compare SkyScript-100M with the existing dataset in detail and demonstrate some deeper insights that can be achieved based on SkyScript-100M. Based on SkyScript-100M, researchers can achieve several deeper and more far-reaching script optimization goals, which may drive a paradigm shift in the entire field of text-to-video and significantly advance the field of short drama video generation. The data and code are available at https://github.com/vaew/SkyScript-100M.