paper-conference

GVCC: Zero-Shot Video Compression via Codebook-Driven Stochastic Rectified Flow
At ultra-low bitrates, high-fidelity reconstruction requires sampling plausible videos from the posterior rather than regressing to …
Ziyue Zeng, Xun Su, Haoyuan Liu, Bingyu Lu, Yui Tatsumi, Hiroshi Watanabe
PDF Cite Project arXiv
Scalable Variable-Rate Video Compression for Humans and Machines via Keyframe Interpolation
Yui Tatsumi, Ziyue Zeng, Hiroshi Watanabe
Cite Project
Flow Residual Segmentation and Generative Reconstruction for Motion-Aware Video Coding
Recent diffusion-based video generation methods have shown strong potential for video compression and frame interpolation; however, …
Ziyue Zeng, Yui Tatsumi, Hiroshi Watanabe
Cite Project DOI
Training-Free Adaptive Quantization for Variable Rate Image Coding for Machines
Image Coding for Machines (ICM) has become increasingly important with the rapid integration of computer vision technology into …
Yui Tatsumi, Ziyue Zeng, Hiroshi Watanabe
PDF Cite Code Project DOI arXiv
Time Step Generating: A Universal Synthesized Deepfake Image Detector
The rise of high-fidelity text-to-image diffusion models has made synthetic images increasingly indistinguishable from real ones, …
Ziyue Zeng, Yupei Guo, Haoyuan Liu, Dingjie Peng, Hiroshi Watanabe
PDF Cite Code Project DOI arXiv
Bidirectional Attention-Gated Motion Injection for Frame Interpolation
We propose Bi-AGMI (Bidirectional Attention-Gated Motion Injection), a lightweight and efficient framework for keyframe interpolation …
Ziyue Zeng, Yui Tatsumi, Hiroshi Watanabe
Cite Project DOI
Seed Selection for Human-Oriented Image Reconstruction via Guided Diffusion
Conventional methods for scalable image coding for humans and machines require the transmission of additional information to achieve …
Yui Tatsumi, Ziyue Zeng, Hiroshi Watanabe
PDF Cite Project DOI arXiv
Explicit Residual-Based Scalable Image Coding for Humans and Machines
Scalable image compression is a technique that progressively reconstructs multiple versions of an image for different requirements. In …
Yui Tatsumi, Ziyue Zeng, Hiroshi Watanabe
PDF Cite Project DOI arXiv
Assessing the Effectiveness of Residual Information in Scalable Image Coding for Humans and Machines
Yui Tatsumi, Ziyue Zeng, Hiroshi Watanabe
Cite Project