8 papers
LoopViT: Scaling Visual ARC with Looped Transformers
Wen-Jie Shu, Xuerui Qiu, Rui-Jie Zhu +3
Recent advances in visual reasoning have leveraged vision transformers to tackle the ARC-AGI benchmark. However, we argue that the feed-forward architecture, where computational de…
OpenSubject: Leveraging Video-Derived Identity and Diversity Priors for Subject-driven Image Generation and Manipulation
Yexin Liu, Manyuan Zhang, Yueze Wang +8
Despite the promising progress in subject-driven image generation, current models often deviate from the reference identities and struggle in complex scenes with multiple subjects.…
TalkVid: A Large-Scale Diversified Dataset for Audio-Driven Talking Head Synthesis
Shunian Chen, Hejin Huang, Yexin Liu +10
Audio-driven talking head synthesis has achieved remarkable photorealism, yet state-of-the-art (SOTA) models exhibit a critical failure: they lack generalization to the full spectr…
Enhancing Vector Quantization with Distributional Matching: A Theoretical and Empirical Study
Xianghong Fang, Litao Guo, Hengchao Chen +8
The success of autoregressive models largely depends on the effectiveness of vector quantization, a technique that discretizes continuous features by mapping them to the nearest co…
When Semantics Mislead Vision: Mitigating Large Multimodal Models Hallucinations in Scene Text Spotting and Understanding
Yan Shu, Hangui Lin, Yexin Liu +7
Large Multimodal Models (LMMs) have achieved impressive progress in visual perception and reasoning. However, when confronted with visually ambiguous or non-semantic scene text, th…
Model Reveals What to Cache: Profiling-Based Feature Reuse for Video Diffusion Models
Xuran Ma, Yexin Liu, Yaofu Liu +5
Recent advances in diffusion models have demonstrated remarkable capabilities in video generation. However, the computational intensity remains a significant challenge for practica…