The International Joint Conference on Artificial Intelligence (IJCAI) 2024, a key event in the AI world, has officially begun. This year, OPPO made a strong impact by showcasing its latest AI advancements and research.
OPPO’s AI Center presented a research paper titled “Zero-shot High-fidelity and Pose-controllable Character Animation” at the conference. This paper introduced PoseAnimate, a new framework designed to tackle common challenges in the Image-to-Video (I2V) field, such as inconsistent character appearances and poor detail retention. PoseAnimate uses pose information to control animation movements while keeping scene content consistent.
Read more: Chardike.com launched Makeup Festival Offers Up to 30% Off on Top Brands
The framework includes three main modules:
- Pose-Aware Control Module (PACM): Ensures precise movement alignment by integrating pose signals.
- Dual Consistency Attention Module (DCAM): Maintains character identity and background details.
- Mask-Guided Decoupling Module (MGDM): Enhances video fidelity by separating characters from the background.
Additionally, Dr. Yang Zhenyu from OPPO AI Center delivered a talk on “Generative AI for Smartphones,” sharing OPPO’s insights on integrating AI technology into smartphones.
Read more: PriyoShop partners with Unilever to Empower MSMEs in Kaliakoir
