-
Important news
-
News
-
In-Depth
-
Shenzhen
-
China
-
World
-
Business
-
Speak Shenzhen
-
Culture
-
Leisure
-
Photos
-
Lifestyle
-
Travel
-
Tech
-
Special Report
-
Digital Paper
-
Opinion
-
Features
-
Kaleidoscope
-
Health
-
Markets
-
Sports
-
Entertainment
-
Business/Markets
-
World Economy
-
Weekend
-
Newsmaker
-
Advertisement
-
Diversions
-
Movies
-
Hotels and Food
-
Yes Teens!
-
News Picks
-
Glamour
-
Campus
-
Budding Writers
-
Fun
-
Qianhai
-
CHTF Special
-
Futian Today
在线翻译:
szdaily -> Tech -> 
Scientists build 3D facial images from DNA sample
    2025-06-06  08:53    Shenzhen Daily

A GROUP of Chinese researchers has developed an innovative technology that allows them to reconstruct three-dimensional models of human faces solely based on the analysis of genetic material. This unique development was recently published in Advanced Science.

The algorithm, called Difface, works by deeply analyzing DNA regions and matching the resulting data with a large database of digital facial models, called point clouds.

Difface first uses a transformer and a spiral convolutional network to match high-dimensional single nucleotide polymorphisms (SNPs) to 3D facial images with low-dimensional features, respectively, while simultaneously establishing a connection between both modalities in latent features, and then uses a diffusion model to reconstruct facial structures from SNP features.

The system was trained using data from over 9,600 Han Chinese, including pairs of genetic information and corresponding 3D facial images.

“Difface uses a pre-trained transformer to extract DNA sequence features, combines it with spiral convolution technology to extract 3D face point cloud features, and further maps multi-modal features to a unified feature space through a comparative learning method, thus achieving accurate generation from DNA sequences to 3D face images based on the diffusion model,” said study co-author Chen Luonan from the Chinese Academy of Sciences.

The tests showed a high level of accuracy in reproducing appearance. Adding information about gender, age, and body type further improved the quality of the result.

Chen, one of the developers of the algorithm, emphasized that the system not only reproduces appearance based on a DNA sample, but also allows for modeling changes in appearance over time.

A potential area of application for this development is forensics, where the technology will significantly increase the efficiency of identifying criminals based on biological traces left at crime scenes.

Besides, this technology may offer a new approach to finding missing children — even if they are now adults, their current appearance can be generated using their DNA. In anthropological research, ancient human DNA might reveal the true faces of our ancestors. In the field of medical aesthetics, it could enable personalized appearance predictions based on genetics, the research team noted.(SD-Agencies)

深圳报业集团版权所有, 未经授权禁止复制; Copyright 2010-2020, All Rights Reserved.
Shenzhen Daily E-mail:szdaily@126.com