Pith. sign in

REVIEW 22 cited by

OpenS2V-Nexus: A Detailed Benchmark and Million-Scale Dataset for Subject-to-Video Generation

Not yet reviewed by Pith; the record is open.

Re-run · record.json Download PDF Read on arXiv ↗

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2505.20292 v4 pith:XUI2Z6Z3 submitted 2025-05-26 cs.CV cs.AI

Shenghai Yuan , Xianyi He , Yufan Deng , Yang Ye , Jinfa Huang , Bin Lin , Jiebo Luo , Li Yuan This is my paper
classification cs.CVcs.AI
keywords generationvideosdatasetopens2v-evalopens2v-nexusbenchmarkbenchmarksbuilding
verification ladder T0 review T1 audit T2 compute T3 formal
0 comments
read the original abstract

Subject-to-Video (S2V) generation aims to create videos that faithfully incorporate reference content, providing enhanced flexibility in the production of videos. To establish the infrastructure for S2V generation, we propose OpenS2V-Nexus, consisting of (i) OpenS2V-Eval, a fine-grained benchmark, and (ii) OpenS2V-5M, a million-scale dataset. In contrast to existing S2V benchmarks inherited from VBench that focus on global and coarse-grained assessment of generated videos, OpenS2V-Eval focuses on the model's ability to generate subject-consistent videos with natural subject appearance and identity fidelity. For these purposes, OpenS2V-Eval introduces 180 prompts from seven major categories of S2V, which incorporate both real and synthetic test data. Furthermore, to accurately align human preferences with S2V benchmarks, we propose three automatic metrics, NexusScore, NaturalScore and GmeScore, to separately quantify subject consistency, naturalness, and text relevance in generated videos. Building on this, we conduct a comprehensive evaluation of 18 representative S2V models, highlighting their strengths and weaknesses across different content. Moreover, we create the first open-source large-scale S2V generation dataset OpenS2V-5M, which consists of five million high-quality 720P subject-text-video triples. Specifically, we ensure subject-information diversity in our dataset by (1) segmenting subjects and building pairing information via cross-video associations and (2) prompting GPT-Image-1 on raw frames to synthesize multi-view representations. Through OpenS2V-Nexus, we deliver a robust infrastructure to accelerate future S2V generation research.

Discussion (0). Sign in to comment.

Forward citations

Cited by 22 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. MSAVBench: Towards Comprehensive and Reliable Evaluation of Multi-Shot Audio-Video Generation

    cs.CV 2026-05 conditional novelty 7.0 of 10

    MSAVBench is the first comprehensive benchmark for multi-shot audio-video generation, spanning video, audio, shot, and reference dimensions with an adaptive evaluation framework that reaches 91.5% Spearman correlation...

  2. MSAVBench: Towards Comprehensive and Reliable Evaluation of Multi-Shot Audio-Video Generation

    cs.CV 2026-05 unverdicted novelty 7.0 of 10

    MSAVBench is the first comprehensive benchmark for multi-shot audio-video generation featuring four dimensions, challenging scenarios, and an adaptive hybrid evaluation framework that achieves 91.5% Spearman correlati...

  3. EntityBench: Towards Entity-Consistent Long-Range Multi-Shot Video Generation

    cs.CV 2026-05 conditional novelty 7.0 of 10

    EntityBench is a new benchmark with detailed per-shot entity schedules from real media, and the EntityMem baseline using persistent per-entity memory achieves the highest character fidelity with Cohen's d of +2.33.

  4. MuSS: A Large-Scale Dataset and Cinematic Narrative Benchmark for Multi-Shot Subject-to-Video Generation

    cs.CV 2026-04 unverdicted novelty 7.0 of 10

    MuSS is a movie-derived dataset and benchmark that enables AI models to generate multi-shot videos with coherent narratives and preserved subject identity across shots.

  5. MuSS: A Large-Scale Dataset and Cinematic Narrative Benchmark for Multi-Shot Subject-to-Video Generation

    cs.CV 2026-04 unverdicted novelty 7.0 of 10

    MuSS is a new movie-sourced dataset and benchmark that enables AI models to generate multi-shot videos with improved narrative coherence and subject identity preservation.

  6. RefCaptioner: Multi-Reference Image-Grounded Video Captioning

    cs.CV 2026-07 conditional novelty 6.0 of 10

    Mixed-data SFT plus Hierarchical Coverage-Discounted GRPO yields open-source SOTA multi-reference image-grounded video captions on MRVBench without hurting general captioning.

  7. Aura: Consistent Multi-Subject Video Generation via VLM-Grounded Semantic Alignment

    cs.CV 2026-07 conditional novelty 6.0 of 10

    Aura combines VLM meta-queries, T5-teacher alignment, subject-aware RoPE shifts, memory tokens, and a large AIGC-curated dataset to claim SOTA multi-element subject-to-video generation under OpenS2V-Eval Total score.

  8. FashionChameleon: Towards Real-Time and Interactive Human-Garment Video Customization

    cs.CV 2026-05 unverdicted novelty 6.0 of 10

    FashionChameleon achieves interactive multi-garment video customization in real time by training a teacher model with in-context learning on single-garment pairs, applying streaming distillation, and using training-fr...

  9. FashionChameleon: Towards Real-Time and Interactive Human-Garment Video Customization

    cs.CV 2026-05 unverdicted novelty 6.0 of 10

    FashionChameleon achieves interactive multi-garment video customization at 23.8 FPS via in-context teacher models, streaming distillation, and training-free KV cache rescheduling while using only single-garment data.

  10. LIVE: Leveraging Image Manipulation Priors for Instruction-based Video Editing

    cs.CV 2026-04 unverdicted novelty 6.0 of 10

    LIVE achieves state-of-the-art instruction-based video editing by jointly training on image and video data with a frame-wise token noise strategy to bridge domain gaps and a new benchmark of over 60 tasks.

  11. OmniShow: Unifying Multimodal Conditions for Human-Object Interaction Video Generation

    cs.CV 2026-04 unverdicted novelty 6.0 of 10

    OmniShow unifies text, image, audio, and pose conditions into an end-to-end model for high-quality human-object interaction video generation and introduces the HOIVG-Bench benchmark, claiming state-of-the-art results.

  12. Rethinking Position Embedding as a Context Controller for Multi-Reference and Multi-Shot Video Generation

    cs.CV 2026-04 conditional novelty 6.0 of 10

    SideInfo-RoPE encodes reference-identity agreement as an extra rotary axis, disambiguating similar characters in multi-reference multi-shot video generation while keeping full semantic attention.

  13. Learning Zero-Shot Subject-Driven Video Generation Using 1% Compute

    cs.CV 2025-04 unverdicted novelty 6.0 of 10

    A zero-shot subject-driven video generation framework that decomposes the task into identity injection from 200K subject-image pairs and motion preservation from 4K arbitrary videos, trained in 288 A100 GPU hours on C...

  14. Vera: Identity-Faithful Human Subject-to-Video Generation

    cs.CV 2026-07 conditional novelty 5.0 of 10

    Vera improves identity consistency in human subject-to-video generation using cross-clip identity-aligned data, face-weighted masked loss, and layer-aware reference attention.

  15. DomainShuttle: Freeform Open Domain Subject-driven Text-to-video Generation

    cs.CV 2026-06 unverdicted novelty 5.0 of 10

    DomainShuttle introduces domain-aware modeling and token separation techniques to achieve high subject fidelity with generative flexibility in open-domain subject-driven text-to-video tasks.

  16. HarmoView: Harmonizing Multi-View Constraints for Identity-Consistent Video Generation

    cs.CV 2026-06 unverdicted novelty 5.0 of 10

    HarmoView proposes Multi-level Feature Injection, learnable proxy tokens, Jump-RoPE, and Progressive View Curriculum plus a new multi-view dataset to achieve state-of-the-art identity-consistent video generation from ...

  17. Bernini: Latent Semantic Planning for Video Diffusion

    cs.CV 2026-05 unverdicted novelty 5.0 of 10

    Bernini is a framework that uses an MLLM planner to output semantic representations for a DiT renderer to generate or edit videos, reporting SOTA benchmark performance.

  18. Omni-Customizer: End-to-End MultiModal Customization for Joint Audio-Video Generation

    cs.CV 2026-05 unverdicted novelty 5.0 of 10

    Omni-Customizer proposes an end-to-end framework using Omni-Context Fusion, Masked TTS Cross-Attention, Semantic-Anchored Multimodal RoPE, and specialized training curricula to achieve precise multimodal identity bind...

  19. Identity-Preserving Text-to-Video Generation via Training-Free Prompt, Image, and Guidance Enhancement

    cs.CV 2025-09 conditional novelty 5.0 of 10

    A training-free prompt, image, and guidance enhancement framework improves face consistency and video quality for identity-preserving text-to-video generation, winning the ACM Multimedia 2025 IPVG challenge.

  20. UniWorld-V1: High-Resolution Semantic Encoders for Unified Visual Understanding and Generation

    cs.CV 2025-06 unverdicted novelty 5.0 of 10

    UniWorld-V1 shows that semantic features from large multimodal models enable unified visual understanding and generation, achieving strong results on perception and manipulation tasks with only 2.7 million training samples.

  21. Collaborative Few-Step Distillation and Low-Bit Quantization for Wan2.2 Dual-Expert Video Diffusion Models

    cs.CV 2026-05 unverdicted novelty 3.0 of 10

    A co-designed few-step distillation and low-bit quantization pipeline for Wan2.2-T2V-A14B keeps quantized few-step performance close to or above the full-precision baseline at 8 and 20 steps.

  22. Tail-Aware HiFloat4: W4A4 Post-Training Quantization for Wan2.2

    cs.AI 2026-05 unverdicted novelty 3.0 of 10

    Adapts ViDiT-Q for W4A4 HiFloat4 quantization of Wan2.2 with tail-aware percentile calibration to limit outlier effects while preserving the original runtime pipeline.

Pith tools