Meta AI has released Sapiens2, a new high-resolution vision model focused on human imagery. This model, built on a unified transformer backbone and trained on 1 billion human images, achieves state-of-the-art performance in tasks like pose estimation, segmentation, and 3D geometry prediction. Sapiens2 also demonstrates strong capabilities in predicting surface normals and albedo. AI
影响 Advances human-centric computer vision capabilities, potentially impacting AR/VR and digital human creation.
排序理由 Release of a new vision model from a major tech company's research division.
在 Mastodon — mastodon.social 阅读 →
AI 生成摘要 · Google Gemini · 来自 7 个来源。 我们如何撰写摘要 →
