Partial least squares recursive craniofacial reconstruction method based on tensor space

Abstract

The invention relates to a method for reconstructing face skin of an unknown skull by conducting partial least squares recursive in a tensor space. The method divides a whole skull and a whole face skin tissue in a training sample into two four-order tensors respectively, the skull and the face skin are transferred to different tensor feature spaces through tensor decomposition, and recursive modeling of face skin tensor features on skull tensor features is conducted according to genders by using the partial least squares technology. For a target skull to be reconstructed, the target skull is transferred to a skull tensor feature space in the same method as the training sample, corresponding face skin tensor features are obtained by using a built recursive model, and a target face skin is reconstructed according to tensor features. The method fully reflects essence relation that the skull decides a face, and an ill-posed problem of lacking solving data in traditional reconstruction based on a statistic model is solved. The method has important application value in the fields of criminal investigation, archaeology, forensic anthropology, medical surgery, facial beautifying plastic surgery and the like.

Description

technical field [0001] The invention relates to a facial skin reconstruction method, in particular to a method for realizing facial skin reconstruction of an unknown skull by performing partial least square regression in a tensor space. It is mainly used in criminal investigation, archaeology, forensic anthropology, medical surgery, cosmetic surgery and other fields. Background technique [0002] The skull is the intrinsic biological feature of the human face, and facial reconstruction based on the shape of the skull is a typical application project in the field of forensic anthropology. The traditional manual reconstruction is based on the undetermined skull, with the help of anatomical measurement and acupuncture measurement to obtain the thickness law of the soft tissue of the face. It is sculpted by an artist and belongs to the field of anthropologists, artists and doctors. The reconstructor's personal grasp of craniofacial features and art is greatly affected by subjec...

AI Technical Summary

Problems solved by technology

This kind of method essentially solves the pathological problem of missing data for craniofacial samples with only a small part of data. The optimization process tends to fall into local minima, and it is difficult to reflect the decisive role of the skull on the facial skin.

Images