A Novel Evolutionary Method for Automated Skull-Face Overlay in Computer-Aided Craniofacial Superimposition

Práxedes Martínez-Moreno; Andrea Valsecchi; Pablo Mesejo; Pilar Navarro-Ramírez; Valentino Lugli; Sergio Damas

cs.CVFeb 26, 2026

A Novel Evolutionary Method for Automated Skull-Face Overlay in Computer-Aided Craniofacial Superimposition

Authors: Práxedes Martínez-Moreno, Andrea Valsecchi, Pablo Mesejo, Pilar Navarro-Ramírez, Valentino Lugli, Sergio Damas

Organizations: Department of Computer Science and Artificial Intelligence, University of Granada, Spain and the Andalusian Research Institute in Data Science and Computational Intelligence, University of Granada, Granada, Spain · Panacea Cooperative Research S. Coop., Ponferrada, Spain · Department of Software Engineering, University of Granada, Spain and the Andalusian Research Institute in Data Science and Computational Intelligence, University of Granada, Granada, Spain

Abstract

Craniofacial Superimposition is a forensic technique for identifying skeletal remains by comparing a post-mortem skull with ante-mortem facial photographs. A critical step in this process is Skull-Face Overlay (SFO). This stage involves aligning a 3D skull model with a 2D facial image, typically guided by cranial and facial landmarks' correspondence. However, its accuracy is undermined by individual variability in soft-tissue thickness, introducing significant uncertainty into the overlay. This paper introduces Lilium, an automated evolutionary method to enhance the accuracy and robustness of SFO. Lilium explicitly models soft-tissue variability using a 3D cone-based representation whose parameters are optimized via a Differential Evolution algorithm. The method enforces anatomical, morphological, and photographic plausibility through a combination of constraints: landmark matching, camera parameter consistency, head pose alignment, skull containment within facial boundaries, and region parallelism. This emulation of the usual forensic practitioners' approach leads Lilium to outperform the state-of-the-art method in terms of both accuracy and robustness.

Figures & tables

Fig. 1: Overview of the CFS process comprising three main stages.

Fig. 2: Cone-based soft-tissue modeling approach for 3D facial landmark estimation. The volume defines an anatomically plausible conical search region with a $40^{\circ}$ angular aperture, originating from a cranial landmark (left). The soft-tissue depth ( $||\vec{ST}_{i}||$ ) is calculated as a linear interpolation between population-based minimum ( $minP_{i}$ ) and maximum ( $maxP_{i}$ ) values using the parameter $p_{1}$ (second from left). The angular aperture ( $\alpha$ ) of the cone is scaled by $p_{2}$ , and the rotation angle ( $\theta$ ) about the cranial vector is controlled by $p_{3}$ , which allows rotation across the entire $360^{\circ}$ range (third and fourth from left, respectively). Together, these parameters define a constrained 3D region where the facial landmark is likely to be found.

Fig. 3: This figure illustrates the Lilium pipeline. The process begins with the random initialization of a solution population, which is then iteratively refined by the DE algorithm through mutation, crossover, and selection. A key element in this refinement is Lilium’s powerful composite fitness function (see Section II-C ). This iterative process continues until the defined stopping criterion is met, at which point the vector with the best fitness score represents the optimal SFO solution.

Fig. 4: Contour extraction for the parallelism term $\text{P}_{\text{pll}}$ . The regions of interest are isolated from the skull and the facial meshes using anatomically defined cutting planes (top). The boundary of each segmented region is extracted, cleaned, ordered, and merged into a single continuous curve. The curves are then trimmed using averaged endpoint normals to ensure consistent and comparable extents between skull and face (bottom).

Fig. 5: Real facial photographs are replaced by synthetic renderings generated from 3D facial models, where landmarks are precisely defined. These 3D landmarks are projected into 2D using known camera parameters, eliminating annotation errors. This visualization is based on a synthetic face model used solely for illustrative purposes. The underlying 3D model is “Female head sculpt” by riceart ( https://skfb.ly/I6rZ ), licensed under Creative Commons Attribution 4.0 ( https://creativecommons.org/licenses/by/4.0/ ).

	Exp A (Ideal)					Exp B (Landm. Noise)					Exp C (Realistic)
	Rank		BPE		Time	Rank		BPE		Time	Rank		BPE		Time
Method	Frontal	Lateral	Frontal	Lateral	(s)	Frontal	Lateral	Frontal	Lateral	(s)	Frontal	Lateral	Frontal	Lateral	(s)
POSEST-SFO	1.18	1.135	1.078	2.158	0.014	1.502	1.424	1.861	2.998	0.018	2.147	2.094	2.328	4.103	0.017
Lilium	1.552	1.883	1.15	1.998	220.286	1.672	1.626	2.296	2.769	222.439	1.909	1.781	2.608	3.311	223.757
Lilium ( $\text{P}_{\text{pll}}$ )	1.632	1.259	1.229	2.223	515.921	1.693	1.257	1.811	2.599	515.432	1.95	1.374	2.174	2.92	512.609
Lilium ( $\text{P}_{\text{skof}}$ )	1.412	1.34	1.084	1.894	386.235	1.625	1.271	2.07	2.539	387.511	1.899	1.341	2.299	2.87	385.442

TABLE I: Mean rank, BPE, and computational time results from Exps. A–C. Bold indicates the best value in each column.

		Rank			BPE (mm)
Exp.	Pose	Lilium	POSEST	p	Lilium	POSEST	p
A	Frontal	1.412	1.18	$<\text{0.0001}$	1.084	1.078	$>0.999$
A	Lateral	1.1	1.135	0.201	1.894	2.158	0.001
B	Frontal	1.576	1.502	$>0.999$	1.679	1.861	$<\text{0.0001}$
B	Lateral	1.111	1.424	$<\text{0.0001}$	2.499	2.998	0.0001
C	Frontal	1.897	2.147	$<\text{0.0001}$	1.967	2.328	$<\text{0.0001}$

TABLE II: Statistical comparison across frontal and lateral poses (Exps. A–C). Significance assessed via Wilcoxon signed-rank test. Bold indicates

p<0.05

and better (lower) values.

Exp.	Method	Implausible SFOs (%)	Avg pixels outside (%)	Positive (n)	Negative (n)
A	POSEST-SFO	70.51	3.98	320	11,902
A	Lilium $\text{P}_{\text{skof}}$	19.47	3.03	18	3,354
B	POSEST-SFO	72.6	4.51	460	12,122
B	Lilium $\text{P}_{\text{skof}}+\text{P}_{\text{pll}}$	17.26	3.18	37	2,952
C	POSEST-SFO	78.29	5.08	634	12,937
C	Lilium $\text{P}_{\text{skof}}+\text{P}_{\text{pll}}$	19.83	2.84	48	3,387

TABLE III: Worst-case implausible SFO metrics across experiments. Reported counts correspond to the number of implausible SFOs out of 17,340 cases. Bold indicates better (lower) values.

Panel of Figure 9 — Fig. 6: Representative SFO results from Exp. C across models for multiple subjects. Each column shows a different model; the last is ground truth. Blue circles mark facial landmarks in the image; green dots are projected 3D estimates.

Cross-Modal Image Registration

Cross-modal image registration aligns images from different sensors or modalities into a shared spatial frame.

See all 70 papers

Cranio-Diff: Diffusion-based Cross-domain Craniofacial Reconstruction with 2D X-ray Skull Guidance and Structural Identity Constraints
Jun 8, 2026Ravi Shankar Prasad, Naresh Gurjar, Shashank Baghel +2Cross-Modal Alignment Conditional Diffusion Models
School of Computing and Electrical Engineering Indian Institute of Technology Mandi Himachal Pradesh, India · University Teaching Department, CSVTU Bhilai, India
Automatic dental superimposition of 3D intraorals and 2D photographs for human identification
Apr 7, 2026Antonio D. Villegas-Yeguas, Xavier Abreau-Freire, Guillermo R-García +52D-3D Image Registration Biometric Identification
Department of Computer Science and Artificial Intelligence of the University of Granada (DECSAI), Spain · Panacea Cooperative Research S. Coop., Ponferrada, Spain · Associate Laboratory i4HB - Institute for Health and Bioeconomy, University Institute of Health Sciences (IUCS), CESPU, 4585-116 Gandra, Portugal +7
SuperFace: Preference-Aligned Facial Expression Estimation Beyond Pseudo Supervision
May 7, 2026Zejian Kang, Xuanyang Xu, Wentao Yang +6Facial Expression Recognition Preference Learning
Zhejiang University · The Chinese University of Hong Kong, Shenzhen · Westlake University +3

Jun 8, 2026J/K move · Enter open · S save

Cranio-Diff: Diffusion-based Cross-domain Craniofacial Reconstruction with 2D X-ray Skull Guidance and Structural Identity Constraints

Ravi Shankar Prasad, Naresh Gurjar, Shashank Baghel, Chirag+1

School of Computing and Electrical Engineering Indian Institute of Technology Mandi Himachal Pradesh, India · University Teaching Department, CSVTU Bhilai, India

Cross-Modal Alignment Conditional Diffusion Models

Open paper AskPDF arXiv

The identification of unknown human remains is a persistent humanitarian and forensic challenge. In forensic investigations, it becomes even more challenging when common biometric DNA identifications, such as soft-tissue or hair strands are not present, as they degrade with time. To narrow down the process of identification, few attempts have been made to identify an individual by reconstructing a face from the given skeletal remains ( i.e., skull) using generative AI models, such as CycleGAN, Pix2Pix, and diffusion models. However, they fail to effectively capture semantic information in craniofacial reconstruction, due to a mismatch in the alignment of structural information across modalities. To address this issue, we propose Cranio-Diff, a diffusion-based framework for craniofacial reconstruction from 2D X-ray skull images. The proposed approach integrates skull-conditioned structural guidance through ControlNet with biometric text prompts to generate a face. The proposed Cranio-Diff method is evaluated on skull-face dataset obtained from X-ray scans of 120 subjects in lateral and frontal views. To enable controlled evaluation, each face image is artificially synthesized across three age groups (25, 45, 65) and three BMI variations 10%, yielding 4320 paired samples. This dataset is also curated across different facial modalities ( RGB and sketch) with different appearance conditions (hair and without hair). To the best of our knowledge, this is the only X-ray-face dataset with this magnitude. Extensive experiments showed that the proposed method achieves improved performance over the recent existing approaches in both generated image quality and retrieval task. Finally, to evaluate the performance of our proposed method, we have evaluated the quality of the generated image using FID, SSIM, LPIPS score.

Abstract

Cross-Modal Image Registration

Explore similar work

Cranio-Diff: Diffusion-based Cross-domain Craniofacial Reconstruction with 2D X-ray Skull Guidance and Structural Identity Constraints

Automatic dental superimposition of 3D intraorals and 2D photographs for human identification

SuperFace: Preference-Aligned Facial Expression Estimation Beyond Pseudo Supervision