Computational Forensics: Digital Skull Restoration and Facial Reconstruction

Project Overview

We study computational forensic skull and face modeling, and develop 3D geometric modeling and analysis tools to facilitate forensic, medical, and archaeological tasks.
Skull Restoration: reassemble fragmented skull evidence and repair their damaged regions.
Skull Assessment: determine the basic bio-profile (ancestry, sex, etc) of the skull.
Facial Reconstruction: recreate craniofacial geometry from the skull.
The restoration and analysis of skull evidence and the reconstruction of unidentified body's face from its skull are important tasks in law enforcement investigation. The current skull and face modeling techniques adoped in law investigation are performed manually by forensic specialists. Despite their success in many cases, the procedures are relatively slow, expensive, and reliant on forensic specialists' expertise.
Our goal is to augment these state-of-the-art manual restoration and reconstruction techniques using novel 3D geometric modeling techniques, and gradually migrate the manual pipeline into a digital environment.

Collaborators

Mary Manhein, Ginesse Listi (LSU Forensic Anthropology and Computer Enhancement Services (FACES) Lab)
Warren Waggenspack (LSU Mechanical Engineering Department)

Data Sources and Digital Skull Database (Currently containing 320+ Skull Data)

Digital Scans of a few Skull Models from Major Ancestries
Digitization of skull models from LSU FACES Lab collection
Louisiana Missing & Unidentified Database
Our digitization (conducted in June 2012) from William Bass Skeletal Collection at University of Tennessee
Our digitization (conducted in July 2012) from Pima County Office of the Medical Examiner (PCOME) in Tucson, Arizona
Our digitization (conducted in July 2012) from Terry Skeletal Collection, Smithsonian National Museum of Natural History.

Skull Digitization and Printing

3D Skull Scanning, Reconstruction, and Printing.

Fragmented or Damaged Skull Restoration

Fragment Reassembly and Damaged-region Restoration.

Ancestry Prediction

Skull Classification and Ancestry Prediction.

Facial Reconstruction

Facial Tissue Modeling and Reconstruction

Selected Publications

C. Liu and X. Li. Superimposition-guided Facial Reconstruction from Skull. arXiv:1810.00107, 2018. [Bibtex] [Paper ]

C. Le and X. Li. JigsawNet: Shredded Image Reassembly using Convolutional Neural Network and Loop-based Composition. arXiv:1809.04137, 2018. [Bibtex] [Paper ] [Codes ]

X. Li, K. Xie, W. Hong, C. Liu. Hierarchical Fragmented Image Reassembly using a Bundle-of-Superpixel Representation. Computer-aided Geometric Design (CAGD)71:220-230, 2019. [Bibtex]

C. Maier, K. Zhang, M. Manhein, and X. Li. Palate Shape and Depth: A Shape-Matching and Machine Learning Method for Estimating Ancestry from Human Skeletal Remains. Journal of Forensic Sciences (JFS), 60(5):1129-1134, 2015. [Bibtex]

K. Zhang, W. Yu, M. Manhein, W. Waggenspack and X. Li. 3D Fragment Reassembly using Integrated Template Guidance and Fracture-Region Matching. International Conference on Computer Vision (ICCV) 2015. [Bibtex]

K. Zhang and X. Li. A Graph-based Optimization Algorithm for Fragmented Image Reassembly. Graphical Models, 76(5):484-498, 2014. [Bibtex]

Z. Yin, L. Wei, M. Manhein, and X. Li. An Automatic Assembly and Completion Framework for Fragmented Skulls. International Conference on Computer Vision (ICCV) 2011. [Bibtex]

W. Yu, M. Li, and X. Li. Fragmented Skull Modeling Using Heat Kernels. Graphical Models, 74(4):140-151, 2012. [Bibtex]

X. Li, Z. Yin, L. Wei, S. Wan, W. Yu, and M. Li. Symmetry and template guided completion of damaged skulls. Computers and Graphics, 35:885-893, 2011. [Bibtex]

L. Wei, W. Yu, M. Li, and X. Li. Skull Assembly and Completion using Template-based Surface Matching. in Proc. International Joint 3DIM/3DPVT Conference: 3D Imaging, Modeling, Processing, Visualization, Transmission (3DIMPVT), pp. 413-420,Hangzhou, China, 2011. [Bibtex]

Geometric Reassembly Benchmark

The following data are released for research purpose.

Please cite our paper [K. Zhang, W. Yu, M. Manhein, W. Waggenspack and X. Li. 3D Fragment Reassembly using Integrated Template Guidance and Fracture-Region Matching. International Conference on Computer Vision (ICCV) 2015. ] [Bibtex]

Some Ceramic Models	Scanned Fragments	Complete Model
Child	Download	Download
Gnome	Download	Download
Chicken	Download	Download
Buddha Head	Download	N/A
Skull Model (19 pieces)	Download	Download

Video from 2015 ICCV paper: "3D Fragment Reassembly using Integrated Template Guidance and Fracture-Region Matching," Kang Zhang, Wuyi Yu, Mary Manhein, Warren Waggenspack, Xin Li, International Conference on Computer Vision (ICCV) 2015.

Video from 2011 ICCV paper: "An Automatic Assembly and Completion Framework for Fragmented Skulls," Zhao Yin, Li Wei, Mary Manhein, Xin Li, International Conference on Computer Vision (ICCV) 2011.