OPPO AI Center’s latest research paper titled ‘Zero-shot High-fidelity and Pose-controllable Character Animation’ has been chosen for the IJCAI 2024, and OPPO has unveiled several advanced AI innovations during the conference; let’s take a look.
In the I2V (Image-to-Video) sector, challenges such as inconsistent character appearances and poor detail retention typically demand extensive video data and significant computational resources.
To tackle these issues, OPPO AI Center developed PoseAnimate, a zero-shot I2V framework that uses pose information to control animation movements while ensuring scene content consistency.
The framework accomplishes this through the following three key modules:
- Pose-Aware Control Module (PACM): Integrates various pose signals into text embeddings to retain content irrelevant to the character and ensure precise alignment of movements.
- Dual Consistency Attention Module (DCAM): Enhances temporal consistency, preserving character identity and intricate background details.
- Mask-Guided Decoupling Module (MGDM): Improves the fidelity of generated videos by decoupling characters from the background.
Furthermore, the OPPO team introduced the Pose Alignment Transition Algorithm (PATA) for enhanced character pose transitions and alignment. Experimental results show that PoseAnimate outperforms existing methods in maintaining character consistency, detail fidelity, and high temporal consistency in animations.
Like our Facebook Page here at NasiLemakTech.com for more news and in-depth reviews! Also, join our Facebook Group for insightful information and memes!
Subscribe to our YouTube channel too!
