Post date: Jul 20, 2012 10:15:17 AM
GRANADA, SPAIN (JULY 3, 2012) (REUTERS) - A new skull recognition technique that could help identify missing victims of mass disasters is being pioneered by Spanish researchers.
A 3D skull scanner that can identify skeletons when decomposition has rendered DNA and fingerprint testing impossible could become a revolutionary new tool in helping recognise victims of mass disasters.
The team based in Granada, southern Spain, have devised Face2Skull, a system based on the forensic identification technique known as craniofacial superimposition. The technique involves analysing the morphology of the face by locating sets of craniometric points on the skull and somatometric points on a photograph of the subject.
Forensic anthropologists use a light-weight hand-held scanner to fire X-rays around the front half of a human skull, helping to build up a complete three-dimensional screen image of the subject's face.
According to study co-author Sergio Damas, principal researcher at the European Centre for Soft Computing, the scanner is designed for easy operation.
"The main aim of this scanner is to acquire an accurate 3D model of the skull in a fast and easy user-friendly fashion, so from that point of view the expert should move around the skull in a comfortable way, in order to precisely scan the whole skull from the different views, in order to acquire the frontal part of the skull," said Damas.
The technique is designed to compliment other techniques, serving to discard potential identities of other missing persons before using more expensive or slower identification techniques, such as DNA analysis. In cases where only a skeleton is remaining it may represent the only chance of recognition.
"DNA analysis or fingerprints are not applicable in many different situations and the only, last, chance to identify a person is the human remains based on the skeleton and in particular the skull of the person, so finally the main aim is to accurately project the skull into the photograph in order for the expert, in order for the forensic anthropologist to make a decision on the forensic identification of that particular person," Damas explained.
Initial tests used Computed Tomography (CAT) scans from live individuals based in the Mediterranean region, the first time that skulls of living subjects were used in such examinations. The researchers created a database using tridimensional cranial coordinates, allowing researchers to determine the spatial relationship between each point on the skull.
Science and Artifical Intelligence professor Oscar Cordon says the technology is simple enough for forensic anthropologists with no image processing experience to use. It allows them to 'clean' images by removing artefacts in the photograph and leaving only the face for identification.
"We have an automatic technique in order that an non-expert person, for instance a forensic anthropologist who has no expertise on image processing is able to do it by himself or herself in a pretty quick and robust way, so they want the model to have an error lower than one millimetre. After that there's some need of some small cleaning in order to remove the artefacts and then have an automatic technique by soft computing methods, so artificial intelligence, that decides within a couple of minutes, between two and four minutes to superimpose the the 3D model of the skull over the 2D face picture," he said.
The study, which was developed from a PhD project by Granada University student Fernando Navarro, could be particularly useful in the aftermath of mass disasters. "For mass disasters, when you only can do a skeleton-based identification you need to go for this kind of soft technique, so even if the sample is corrupted and there are not soft tissues, here you only need some preliminary information from the relatives, so the face, at least one face picture, and then the skull you work with and even with a damaged skull you can apply the technique," said Cordon.
Face2Skull, which its creators say can recognise skulls in at least 85 percent of cases, was developed by a multidisciplinary team including two institutions - the European Centre for Soft Computing (ECSC) and the University of Granada who have a joint international patent on the software system. Police organisations in various countries have expressed an interest. A contract to supply the software has already been signed with the national police force of Mexico.
It's already been used by the Spanish Guardia Civil to identify the remains of an elderly Portuguese man whose body was found in an advanced state of decomposition in Granada.