Carmen Campomanes-Alvarez

Ground truth data generation for skull–face overlay

Objective and unbiased validation studies over a significant number of cases are required to get a more solid picture on craniofacial superimposition reliability. It will not be possible to compare the performance of existing and upcoming methods for craniofacial superimposition without a common forensic database available for the research community. Skull–face overlay is a key task within craniofacial superimposition that has a direct influence on the subsequent task devoted to evaluate the skull–face relationships. In this work, we present the procedure to create for the first time such a dataset. We have also created a database with 19 skull–face overlay cases for which we are trying to overcome legal issues that allow us to make it public. The quantitative analysis made in the segmentation and registration stages, together with the visual assessment of the 19 face-to-face overlays, allows us to conclude that the results can be considered as a gold standard. With such a ground truth dataset, a new horizon is opened for the development of new automatic methods whose performance could be now objectively measured and compared against previous and future proposals. Additionally, other uses are expected to be explored to better understand the visual evaluation process of craniofacial relationships in craniofacial identification. It could be very useful also as a starting point for further studies on the prediction of the resulting facial morphology after corrective or reconstructive interventionism in maxillofacial surgery.

Modeling Facial Soft Tissue Thickness for Automatic Skull-Face Overlay

Craniofacial superimposition involves the process of overlaying a skull with a number of ante-mortem images of an individual and the analysis of their morphological correspondence. Within the craniofacial superimposition process, the skull-face overlay stage focuses on achieving the best possible overlay of the skull and a single ante-mortem image of a missing person. This technique has been commonly applied following a nonautomatic trial-and-error approach. Automatic skull-face overlay methods have been developed obtaining promising results. In this paper, we present two new variants that are an extension of existing 3-D-2-D methods to automatically superimpose a skull 3-D model on a facial photograph. We have modeled the imprecision related to the facial soft tissue depth between corresponding pairs of cranial and facial landmarks which typically guide the automatic approaches. As an illustration of the models performance, the soft tissue distances associated to studies for Mediterranean population have been considered for dealing with this landmark matching uncertainty. Hence, we directly incorporate the corresponding landmark spatial relationships within the automatic skull-face overlay procedure. We have tested the performance of our proposal on 18 skull-face overlay instances from a ground truth data set obtaining valuable results. The current proposal is thus the first automatic skull-face overlay method evaluated in a reliable and unbiased way.

Design of criteria to assess craniofacial correspondence in forensic identification based on computer vision and fuzzy integrals

Graphical abstractDisplay Omitted Craniofacial superimposition is one of the most relevant skeleton-based identification techniques. Within this process, the skull-face overlay stage focuses on achieving the best possible overlay of a skull found and an ante mortem image of a candidate person. In previous work, we proposed an automatic skull-face overlay method, based on evolutionary algorithms and fuzzy sets. The following stage, decision making, consists of determining the degree of support of being the same person or not. This decision is based on the analysis of some criteria assessing the skull-face morphological correspondence through the resulting skull-face overlay. In this work, we take a first step to design a decision support system for craniofacial superimposition. To do so, we consider the modeling of two of the most discriminative criteria for assessing craniofacial correspondence: the morphological and spatial relationship between the bony and facial chin, and the relative position of the orbits and the eyeballs. For each criterion, different computer vision-based approaches have been studied. The accuracy of each method has been calculated as its capability to discriminate in a cross-comparison identification scenario. Sugeno integral has been used to aggregate the results of the different methods taking into account the corresponding individual accuracy index. This allows us to provide a single global output specifying the matching of each criterion while combining the capabilities of different methods. Finally, the performance of the designed criteria and methods have been tested on 172 skull-face overlay problem instances of positive and negative cases to illustrate the discriminative power of each criterion. It has been shown that thanks to the use of Sugeno integral for aggregating different methods, a more robust measurement output is achieved.

Study on the criteria for assessing skull-face correspondence in craniofacial superimposition

Craniofacial superimposition has the potential to be used as an identification method when other traditional biological techniques are not applicable due to insufficient quality or absence of ante-mortem and post-mortem data. Despite having been used in many countries as a method of inclusion and exclusion for over a century it lacks standards. Thus, the purpose of this research is to provide forensic practitioners with standard criteria for analysing skull-face relationships. Thirty-seven experts from 16 different institutions participated in this study, which consisted of evaluating 65 criteria for assessing skull-face anatomical consistency on a sample of 24 different skull-face superimpositions. An unbiased statistical analysis established the most objective and discriminative criteria. Results did not show strong associations, however, important insights to address lack of standards were provided. In addition, a novel methodology for understanding and standardizing identification methods based on the observation of morphological patterns has been proposed.

Hierarchical information fusion for decision making in craniofacial superimposition

Craniofacial superimposition is one of the most important skeleton-based identification methods. The process studies the possible correspondence between a found skull and a candidate (missing person) through the superimposition of the former over a variable number of images of the face of the latter. Within craniofacial superimposition we identified three different stages, namely: (1) image acquisition-processing and landmark location; (2) skull-face overlay; and (3) decision making. While we have already proposed and validated an automatic skull-face overlay technique in previous works, the final identification stage, decision making, is still performed manually by the expert. This consists of the determination of the degree of support for the assertion that the skull and the ante-mortem image belong to the same person. This decision is made through the analysis of several criteria assessing the skull-face anatomical correspondence based on the resulting skull-face overlay. In this contribution, we present a hierarchical framework for information fusion to support the anthropologist expert in the decision making stage. The main goal is the automation of this stage based on the use of several skull-face anatomical criteria combined at different levels by means of fuzzy aggregation functions. We have implemented two different experiments for our framework. The first aims to obtain the most suitable aggregation functions for the system and the second validates the proposed framework as an identification system. We tested the framework with a dataset of 33 positive and 411 negative identification instances. The present proposal is the first automatic craniofacial superimposition decision support system evaluated in an objective and statistically meaningful way.

Experimental study of different aggregation functions for modeling craniofacial correspondence in craniofacial superimposition

Craniofacial superimposition is a forensic identification method involving the overlay of a skull over the available ante-mortem photographs of a candidate missing person face and the subsequent analysis of their anatomical correspondence. Within this process, the decision making stage focuses on determining the degree of support of being the same person or not based on the analysis of some criteria assessing the skull-face morphological correspondence. That decision is usually made in a non automatic and subjective way. We aim to automate the decision making process using computer vision and soft computing methods to assist the forensic anthropologist. In a previous study we have developed several methods to measure the matching of the correspondence between the face and the skull. The accuracy of each method was calculated as its capability to discriminate in a cross-comparison identification scenario. By the use of aggregation functions we can combine the results of the different methods taking into account the corresponding individual accuracy. This allows us to provide a single global output specifying the matching of each criterion while combining the capability of different methods. In this work, we present a study of the behavior of different aggregation functions for this aim. The performance of the aggregated methods has been tested on 172 skull-face overlay problem instances of positive and negative cases. The obtained results show that Sugeno integral ranks better than the counterparts although not significant conclusions can be delivered regarding the performance.

Genetic algorithms for skull-face overlay including mandible articulation

Abstract Craniofacial superimposition (CFS) is a skeleton-based technique that aims to provide identity to a skull through its superimposition with one or more photographs of candidate missing people. While traditionally performed by forensic experts, computer-aided CFS methods can now provide substantial speedups and are quickly progressing towards a large degree of automation. A current major limitation concerns the position of the mandible, which is required to be manually set by the expert beforehand in order to reproduce the facial expression of the subject in each available photograph. This is time-consuming and prone to errors. In this work, we address this issue by extending the state-of-the-art genetic algorithm-based method with the ability to allocate the mandible in the right position according to an anatomical model. Based on a dataset of simulated ante-mortem images with different mandible apertures and facial poses, we prove experimentally that the proposed method is able to effectively tackle cases displaying a much larger range of mandible positions. In fact, thanks to the new genetic design, it is able to outperform the original method, even when the mandible aperture is very small.