RealDPO: Real or Not Real, that is the Preference
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2025
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866909888864583680 |
|---|---|
| author | Cheng, Guo Yang, Danni Huang, Ziqi Si, Jianlou Si, Chenyang Liu, Ziwei |
| author_facet | Cheng, Guo Yang, Danni Huang, Ziqi Si, Jianlou Si, Chenyang Liu, Ziwei |
| contents | Video generative models have recently achieved notable advancements in synthesis quality. However, generating complex motions remains a critical challenge, as existing models often struggle to produce natural, smooth, and contextually consistent movements. This gap between generated and real-world motions limits their practical applicability. To address this issue, we introduce RealDPO, a novel alignment paradigm that leverages real-world data as positive samples for preference learning, enabling more accurate motion synthesis. Unlike traditional supervised fine-tuning (SFT), which offers limited corrective feedback, RealDPO employs Direct Preference Optimization (DPO) with a tailored loss function to enhance motion realism. By contrasting real-world videos with erroneous model outputs, RealDPO enables iterative self-correction, progressively refining motion quality. To support post-training in complex motion synthesis, we propose RealAction-5K, a curated dataset of high-quality videos capturing human daily activities with rich and precise motion details. Extensive experiments demonstrate that RealDPO significantly improves video quality, text alignment, and motion realism compared to state-of-the-art models and existing preference optimization techniques. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2510_14955 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | RealDPO: Real or Not Real, that is the Preference Cheng, Guo Yang, Danni Huang, Ziqi Si, Jianlou Si, Chenyang Liu, Ziwei Computer Vision and Pattern Recognition Artificial Intelligence Video generative models have recently achieved notable advancements in synthesis quality. However, generating complex motions remains a critical challenge, as existing models often struggle to produce natural, smooth, and contextually consistent movements. This gap between generated and real-world motions limits their practical applicability. To address this issue, we introduce RealDPO, a novel alignment paradigm that leverages real-world data as positive samples for preference learning, enabling more accurate motion synthesis. Unlike traditional supervised fine-tuning (SFT), which offers limited corrective feedback, RealDPO employs Direct Preference Optimization (DPO) with a tailored loss function to enhance motion realism. By contrasting real-world videos with erroneous model outputs, RealDPO enables iterative self-correction, progressively refining motion quality. To support post-training in complex motion synthesis, we propose RealAction-5K, a curated dataset of high-quality videos capturing human daily activities with rich and precise motion details. Extensive experiments demonstrate that RealDPO significantly improves video quality, text alignment, and motion realism compared to state-of-the-art models and existing preference optimization techniques. |
| title | RealDPO: Real or Not Real, that is the Preference |
| topic | Computer Vision and Pattern Recognition Artificial Intelligence |
| url | https://arxiv.org/abs/2510.14955 |