背景人群按视角重投影
Background figures follow each view
分类：通用视觉生成/编辑与一致性
提示词
同一场景的背景人物位置按提供关系在各视图中对应，远近尺度随机位变化，不把相同尺寸人群粘贴到每张图。
Keep background people at corresponding locations across views using the supplied relationships, adjusting apparent scale to each camera rather than pasting an equally sized crowd into every image.
标签：通用视觉生成、保留与纠错
进入 AI 创作