Periodic Reporting for period 1 - UMAFAE (Unaccompanied Minors Automatic Forensic Age Estimation)
Reporting period: 2022-03-01 to 2024-02-29
UMAFAE’s goal is to develop Artificial Intelligence (AI)-based algorithms that estimate the age of a migrant minor without any family reference by analysing radiographic images. Age is an innate characteristic of our own identity. However, in many countries, especially on the African continent, less than 10% of new births have been reported in a civil registry, in contrast to the official rate in European countries (90%). For this reason, applicants for international protection may arrive in the EU undocumented, or with unreliable identity documents. As a more serious consequence, these minors may end up deprived of the rights to which they are entitled.
Therefore, UMAFAE is of extreme relevance due to its social and legal implications. It provides a key contribution to the migration challenge through more efficient systems that will support forensic experts in the estimation process, contributing to the “Inclusive, Innovative and Thoughtful Societies” challenge. Additionally, UMAFAE is completely aligned with the European Union’s 17 sustainable development goals, specifically, Goal 16 (Peace, Justice and Strong Institutions), since it promotes inclusive societies and build effective and capable institutions at all levels.
The specific objectives are as follows: 1) to improve the accuracy of current age estimation methods in minors by means of AI-based techniques; 2) to develop fully automated methods; 3) to shift from current observational methods to objective, fast, accurate, robust and reproducible ones; 4) to develop models that are explainable and interpretable for the forensic community, especially for the courts of justice, responsible for the decision-making process; 5) to validate the models on real cases of migrant minors without family references; and 6) to contribute to the EU’s action and agenda to protect the rights of minors. Regarding point 5, carrying out the validation on real cases of undocumented migrant minors, whose files are stored in the most important institutes of forensic medicine and forensic sciences in Spain, is a necessary step to demonstrate the validity of the data, the robustness of the algorithms and the reliability of the process itself. The collaboration of the indicated institutions allows us to close a cycle, and to find out whether forensic professionals will really be able to apply these models in their forensic work, according to sound statistical standards and the most important principles of child protection.
Therefore, UMAFAE is of extreme relevance due to its social and legal implications. It provides a key contribution to the migration challenge through more efficient systems that will support forensic experts in the estimation process, contributing to the “Inclusive, Innovative and Thoughtful Societies” challenge. Additionally, UMAFAE is completely aligned with the European Union’s 17 sustainable development goals, specifically, Goal 16 (Peace, Justice and Strong Institutions), since it promotes inclusive societies and build effective and capable institutions at all levels.
The specific objectives are as follows: 1) to improve the accuracy of current age estimation methods in minors by means of AI-based techniques; 2) to develop fully automated methods; 3) to shift from current observational methods to objective, fast, accurate, robust and reproducible ones; 4) to develop models that are explainable and interpretable for the forensic community, especially for the courts of justice, responsible for the decision-making process; 5) to validate the models on real cases of migrant minors without family references; and 6) to contribute to the EU’s action and agenda to protect the rights of minors. Regarding point 5, carrying out the validation on real cases of undocumented migrant minors, whose files are stored in the most important institutes of forensic medicine and forensic sciences in Spain, is a necessary step to demonstrate the validity of the data, the robustness of the algorithms and the reliability of the process itself. The collaboration of the indicated institutions allows us to close a cycle, and to find out whether forensic professionals will really be able to apply these models in their forensic work, according to sound statistical standards and the most important principles of child protection.
A unique global sample has been collected, including thousands of panoramic radiographs (OPT), cone beam computed tomography (CBCT) of the maxillofacial region, hand-wrist X-rays, and chest X-rays or computed tomographies (CT). A sample of more than 10,000 OPTs (10,739) was collected from universities and private dental clinics in more than 20 countries, from subjects of known sex, aged 14-26 years. In addition, medical images stored in public databases including whole body CT scans (New Mexico Decedent Image Database), chest X-rays (Clinical Center – American Research Hospital) and carpal X-rays (Radiological Society of North America – Pediatric Bone Age Challenge) were used. All samples are stored in a highly secure server of the University of Granada (UGR).
Regarding the methodology used for the analysis of the OPTs, different deep learning techniques were applied. In all cases, the images were preprocessed to obtain uniform images in terms of characteristics, regardless of the geographical origin of each sample. Then, different experiments were carried out with convolutional neural networks analyzing both the entire OPT and the third molar only. Furthermore, sex and geographical origin of the study subject has been considered. Finally, the study sample was divided into three parts, training (60%), validation (20%) and test (20%). The best model gave us an error of 1.12 years, with an accuracy of 88.38%. In the case of the analysis of carpal radiographs, an error of 0.57 years was obtained in a sample of subjects aged between 14 and 19 years.
Currently, a validation process is being carried out between human experts and artificial intelligence through the analysis of OPT of real cases of suspected minors from Morocco. Researchers with different expertise have been involved in this process. The aim is to test the efficiency of the models developed in real forensic practice situations and to find out the agreement between different types of observers, human and machine. In addition, a parallel study of three-dimensional images (CBCT and CT) is being carried out to automatically detect, from among the different slices, the sections corresponding to the maxillofacial region and the thorax, i.e. the anatomical areas of most interest for legal age estimation purposes: 4 1) the teeth; and 2) the clavicle. Once the process is automated, artificial neural networks can be applied to determine the accuracy and efficiency of the models in estimating legal age.
Special mention should be made of the section on dissemination, and the section on technology transfer. Apart from the national and international congresses at which the results have been presented (Tenerife, Alicante, Madrid, Bogota, Toulouse, Termoli, San José de Costa Rica, Santiago de Chile, Sydney, Denver), various activities have been organised at schools, institutes, universities and for the general public. These activities have consisted of research seminars, workshops, talks and guest lectures. In addition, Stefano De Luca participated in the “Researchers’ Night” in the city of Granada (Spain). With regard to technology transfer activities, a patent application on age estimation has been filed at European level.
Regarding the methodology used for the analysis of the OPTs, different deep learning techniques were applied. In all cases, the images were preprocessed to obtain uniform images in terms of characteristics, regardless of the geographical origin of each sample. Then, different experiments were carried out with convolutional neural networks analyzing both the entire OPT and the third molar only. Furthermore, sex and geographical origin of the study subject has been considered. Finally, the study sample was divided into three parts, training (60%), validation (20%) and test (20%). The best model gave us an error of 1.12 years, with an accuracy of 88.38%. In the case of the analysis of carpal radiographs, an error of 0.57 years was obtained in a sample of subjects aged between 14 and 19 years.
Currently, a validation process is being carried out between human experts and artificial intelligence through the analysis of OPT of real cases of suspected minors from Morocco. Researchers with different expertise have been involved in this process. The aim is to test the efficiency of the models developed in real forensic practice situations and to find out the agreement between different types of observers, human and machine. In addition, a parallel study of three-dimensional images (CBCT and CT) is being carried out to automatically detect, from among the different slices, the sections corresponding to the maxillofacial region and the thorax, i.e. the anatomical areas of most interest for legal age estimation purposes: 4 1) the teeth; and 2) the clavicle. Once the process is automated, artificial neural networks can be applied to determine the accuracy and efficiency of the models in estimating legal age.
Special mention should be made of the section on dissemination, and the section on technology transfer. Apart from the national and international congresses at which the results have been presented (Tenerife, Alicante, Madrid, Bogota, Toulouse, Termoli, San José de Costa Rica, Santiago de Chile, Sydney, Denver), various activities have been organised at schools, institutes, universities and for the general public. These activities have consisted of research seminars, workshops, talks and guest lectures. In addition, Stefano De Luca participated in the “Researchers’ Night” in the city of Granada (Spain). With regard to technology transfer activities, a patent application on age estimation has been filed at European level.
The main obtained results are represented by a novel age estimation (AE) paradigm, the scientific and experimental outcomes already achieved, and a research prototype. The main exploitation scenario consists of turning the latter demonstrator into a commercial product. Such a product will provide forensic teams with a valuable tool and asset for daily practice. It will reduce costs and AE time while bringing accuracy and reliability to the AE process. In addition, a commercial solution including UMAFAE’s technology will help governments and professionals in the EU preserve the Human Rights included in the EU Charter of Fundamental Rights.
Another key objective of the exploitation strategy is to draw the attention of the corresponding international and national authorities to the need of this research and its ultimate benefits, including generating political and social demand for the methods and product to be developed. The International Committee of Red Cross (ICRD), European Union Agency for Asylum (EUAA), the Latin-Americans Forensic teams, some NGOs (Amnesty International, Andalucía Acoge), and civil society associations (Girls not brides) are especially important supporting organisations, due to their direct relation with advisers of children and adolescents, diplomatic and consular channels and social sectors. Usually, they are involved in the whole ID process so the ER has provided them with information about the basics of the methodology developed, and he has responded to their questions and suggestions. The ER has agreed with these associations to participate in their regular meetings, as well as to organize communication activities together, specially with the ICRD.
Another key objective of the exploitation strategy is to draw the attention of the corresponding international and national authorities to the need of this research and its ultimate benefits, including generating political and social demand for the methods and product to be developed. The International Committee of Red Cross (ICRD), European Union Agency for Asylum (EUAA), the Latin-Americans Forensic teams, some NGOs (Amnesty International, Andalucía Acoge), and civil society associations (Girls not brides) are especially important supporting organisations, due to their direct relation with advisers of children and adolescents, diplomatic and consular channels and social sectors. Usually, they are involved in the whole ID process so the ER has provided them with information about the basics of the methodology developed, and he has responded to their questions and suggestions. The ER has agreed with these associations to participate in their regular meetings, as well as to organize communication activities together, specially with the ICRD.