阅读说明：本技术 用于渲染图像的三维建模体积 (Three-dimensional modeled volume for rendering images ) 是由 斯蒂芬·安东尼·隆巴迪 詹森·萨拉吉 托马斯·西蒙克罗伊茨 加布里埃尔·百洛维茨·施瓦兹 于 2020-03-31 设计创作，主要内容包括：在一个实施例中,一种方法包括：访问对象的多个图片；通过使用机器学习框架处理多个图片来构建用于对象的三维建模的建模体积,其中建模体积与多个颜色和不透明度信息相关联,这些颜色和不透明度信息与建模体积中的多个区域相关联；以及使用建模体积从视点渲染对象的图像,其中通过以下操作渲染图像的每个像素：从视点投射虚拟射线并穿过建模体积,确定建模体积中的多个区域中的与虚拟射线相交的一个或更多个区域,并基于与多个区域中的和虚拟射线相交的一个或更多个区域相关联的颜色和不透明度信息的累积来确定像素的颜色和不透明度。(In one embodiment, a method comprises: accessing a plurality of pictures of an object; constructing a modeled volume for three-dimensional modeling of an object by processing a plurality of pictures using a machine learning framework, wherein the modeled volume is associated with a plurality of color and opacity information associated with a plurality of regions in the modeled volume; and rendering an image of the object from the viewpoint using the modeled volume, wherein each pixel of the image is rendered by: the method includes casting a virtual ray from a viewpoint and through a modeled volume, determining one or more of a plurality of regions in the modeled volume that intersect the virtual ray, and determining a color and opacity of a pixel based on an accumulation of color and opacity information associated with the one or more of the plurality of regions that intersect the virtual ray.)