Dynamic human body three-dimensional model sequence compression method, electronic equipment and medium

Abstract

The invention relates to the field of data compression, and discloses a compression method of a dynamic human body three-dimensional model sequence, electronic equipment and a medium, and the compression method comprises the following steps: dividing the three-dimensional model sequence into a plurality of model sequence segments, and selecting a first KF for each model sequence segment; encrypting the face grid of the first KF model to enable the density of the face grid to be greater than that of the body part grid; then collecting SED-nod for the face grid and the body part grid to form a complete SED-graph, wherein the distance between the face grid collection points is smaller than the distance between the body part grid collection points; performing blocked non-rigid motion solutionon the first KF model to realize non-rigid deformation in the front and rear frame directions; and keeping the texture coordinates of the triangular surface of the model in each model sequence segmentunchanged, and mapping the deformed model by using the original picture. According to the method, the data volume of the model sequence can be greatly compressed, the problem of serious jitter is solved, and the model quality is optimized.

Description

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine

Login to View More

AI Technical Summary

Benefits of technology

The technology allows for better compression or reduction of large amounts of data when created by 3D models without requiring any particular conditions like their position during use. This results in improved accuracy over time compared to traditional methods that require strict alignment between different parts of the body being modeled.

Problems solved by technology

This patented describes how high precision digital images (images) from multiple cameras or sensors used during virtual worlds may contain significant amounts of redundant detail that needs to be removed before being displayed onto an animated screen. However, current methods such as image processing software often result in incorrect results when decompressing this type of dataset due to differences between these two types of datasets. Additionally, existing techniques like depth estimation algorithms require extensive computational resources and slow down their speed because they use complex mathematical equations involving many variables including light intensity levels, geometry features, etc., making it challenging to achieve realistically accurate animation without overwhelming any computing power.

Images