Modeling the consistency between the bony and facial chin outline in craniofacial superimposition

Abstract

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 obtained 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 this work, we present a first approach to model one of the criteria followed by forensic experts: the analysis of the consistency of bony and facial chin outlines. We show some preliminary results over 82 skull-face overlay instances and discuss future research directions.

Modeling Skull-Face Anatomical/Morphological Correspondence for Craniofacial Superimposition-Based Identification

“…To do so, we aim to model the former four families of criteria using computer vision (CV) and soft computing (SC) techniques. CV was already used in our previous works [35], [37]- [39] to measure some craniofacial correspondences and segment the correct contours of the regions. Fuzzy integrals [40], wellknown aggregation operators from the SC field, were used to obtain the final matching degree from several CV methods.…”

“…In [35], [37], [39], we presented two of the most discriminative criteria to assess craniofacial correspondence at criterion evaluation: the morphological and spatial relationship between the facial and bony chin, and the relative position of the eyeballs and the orbits. In these studies, we developed several CV-based methods to assess the degree of matching of each of these two criteria, and aggregated their results in a single value (using different aggregation functions) to obtain more robust and accurate results (level 3 of the DSS framework).…”

“…In previous work [35], we designed some methods to analyze the craniofacial correspondence based on some specific regions. In the current contribution, we study and model a set of suitable criteria to evaluate the morphological/anatomical correspondence on the whole face.…”

“…3) New cases of study using the first three families of criteria to assess craniofacial correspondence: In previous works [35], [38], we presented three cases of study for analyzing the correspondence between the skull and the face in a SFO using CV: the chin outlines comparison, the orbits and eyeball centers positional relations, and the cranial contours comparison. In real cases, experts consider more criteria to establish the craniofacial correspondence.…”

“…This way, we have studied and modeled landmark location and matching uncertainties [33]- [35], and proposed a mathematical formulation [24] and several optimization algorithms for searching the acquisition parameters of the photographs leading to the optimal skull-face overlay. Our system relies on the use of 3D (surface) skull models.…”