Xiaomi and Wuhan University Release WorldCast for Large-Angle Camera-Controlled Character Video
Lei Cao · hf · 2026-10-27
WorldCast from Wuhan University and Xiaomi's MiLM Plus tackles precise camera control in cinematic character video generation. Its CLACT method combines 3D point-cloud guidance with geometry-aligned memory attention to keep characters world-consistent under large-angle trajectories. Project page, code, dataset and model card are public.
More from Multimodal
- Kling 4 generates 30-second drone-style footage, full prompt included — TheLocalLab · 2026-10-27
- Creator has Claude edit her iPhone Duo review inside Tella's agent video editor — evielync · 2026-10-27
- Qwen-Image-PE: two-stage training lifts text-to-image prompts by 7.81 points — Qwen · 2026-10-27
- Curated list of 20 open-source motion libraries for building an AI motion studio — nikola_mr64990 · 2026-10-27
- Redditor builds AI fashion influencer explaining capsule wardrobes — can you tell? — Ai-Emiliaa344 · 2026-10-27
- MOSS Image Edit vs Krea 2 Edit: same prompt side-by-side shirt-swap test — Illustrious_Row_9971 · 2026-10-27