System for automatically generating three-dimensional figure of picture for game

Abstract

The invention discloses a system for automatically generating a three-dimensional figure of a picture for a game. The system comprises an image input and analysis module, an automatic characteristic point generation module, a characteristic point adjustment module, a skeleton system adjustment module, a covering application module, a picture block generation module, a block fusion module and an application module, wherein the image input and analysis module, the automatic characteristic point generation module, the characteristic point adjustment module and the skeleton system adjustment module are connected in turn; an output end of the skeleton system adjustment module is connected with input ends of the covering application module and the picture block generation module in parallel; the output end of the picture block generation module is connected with the input end of the block fusion module; and the input end of the application module is connected with the output ends of the covering application module and the block fusion module in parallel. The system can automatically generate a characteristic three-dimensional model head portrait which meets a front face image, and the actual picture and the block of a model are fused naturally, so that natural face animation can be performed, and the system has low cost, is simple to operate and is particularly suitable for game making, computer graphic (CG) making and three-dimensional music video (MV) making.

Description

technical field [0001] The invention relates to a system for generating three-dimensional graphics from photos, in particular to a system for automatically generating three-dimensional graphics from photos used in games. Background technique [0002] Skin application is a production technique of 3D animation. Add bones to the model based on the model created in the 3D software. Since the skeleton and the model are independent of each other, in order for the skeleton to drive the model to produce reasonable motion. The technique of binding a model to bones is called skinning. [0003] At present, a prior art method for building a 3D face model that integrates multi-view and multi-cue two-dimensional information, first uses two calibrated upper and lower stereo cameras to shoot the face of the modeling object, and obtains the expressionless expression of the face. When changing from the front to the side of the video sequence; then use the fully automatic attitude estimatio...

Images