| CPC G06T 17/00 (2013.01) [G06T 2210/41 (2013.01)] | 19 Claims |

|
1. A method for generating a three-dimensional (3D) image implemented on a computing device having at least one processor and at least one storage device, the method comprising:
obtaining a target two-dimensional (2D) image of a subject;
obtaining a 3D image generation model that includes a plurality of layers, wherein each layer of at least two layers of the plurality of layers includes a down-sampling block, a connection block, and an up-sampling block; and
determining a target 3D image of the subject by processing the target 2D image using the 3D image generation model,
wherein during the processing of the target 2D image, for each layer of the at least two layers,
the down-sampling block is configured to generate a plurality of 2D feature maps by coding an input of the down-sampling block, the input of the down-sampling block including an output generated by a previous layer of the layer based on the target 2D image,
the connection block is configured to generate, based on the plurality of 2D feature maps, at least one 3D feature map; and
the up-sampling block is configured to generate at least one target 3D feature map associated with the target 3D image by processing the at least one 3D feature map,
wherein the obtaining the 3D image generation model further includes:
determining a posture of the subject based on the target 2D image of the subject, the posture of the subject reflecting one or more of a position, a pose, or a shape of the subject in the target 2D image;
selecting, from a model library that includes a plurality of candidate 3D image generation models corresponding to a plurality of postures of the subject, a candidate 3D image generation model corresponding to the posture of the subject; and
designating the selected candidate 3D image generation model as the 3D image generation model.
|