New method for height estimation of subjects represented in photograms taken from video surveillance systems

Evaluation of the effect of orthodontic treatment on the reliability of facial recognition by using three-dimensional model superimposition technique

3D-3D registration of facial models, has great advantages in personal individual identification in forensic medicine. However, orthodontic treatment has brought changes in facial soft and hard tissues, which has a potential effect in personal identification. The aim of the study was to explore whether orthodontic treatment affects 3D-3D facial recognition. A total of 68 patients aged between 18 and 38 were selected (30 subjects with tooth extraction, 38 subjects with non-tooth extraction) and a control group consisting of 30 volunteers without orthodontic treatment was selected from a database of 3dMD facial models. 3dMD facial model acquisition was performed twice for each subject. For both extraction and non-extraction groups, T0 was acquired before treatment, and T1 was acquired at the end of treatment. The time span of 2 times 3dMD facial model acquisition in the control group was about 2 years. 3dMD facial models were then registered onto other 3dMD models belonging to the same and different individuals according to the minimum point-to-point distance, getting a cohort of matches and mismatches. Root mean square (RMS) value of the minimum point-to-point distance between two models was then calculated. The intra- and inter-observer repeatability coefficients were 0.986 (P<0.001) and 0.982 (P<0.001). The absolute technical error of measurement (TEM) value was 0.05 mm and 0.04 mm, and relative technical error of measurement (rTEM) value was 4.91 % and 3.60 %, respectively. Possible significant differences between groups were assessed through independent samples t-test or Mann-Whitney U test (p<0.001). The average RMS value was 1.03±0.33 mm in matches of orthodontic group, 0.67±0.22 mm in matches of control group, and 2.66±0.51 mm in mismatches. An RMS value range of 1.70-1.77 mm could distinguish matches from mismatches in 100 % of cases in the present study. This study showed that orthodontic treatment would not affect 3D-3D facial recognition in adults for the purpose of individual identification in forensic medicine.

Reconsideration of Bertillonage in the age of digitalisation: Digital anthropometric patterns as a promising method for establishing identity

The idea of using measurements of the human body for identity matching is deeply associated with Bertillonage, a historic biometric system that was briefly applied until it was superseded by fingerprinting in the early 20th century. The apparent failure then commonly causes doubt with regard to the suitability of a set of measurements as a biometric identifier in the present. Hence, the aim of this paper is to explore the potentials of using an anthropometric pattern, comprising of a set of body measurements, for identity matching. For this purpose, it will begin with a thorough examination of Bertillon's system and move on to conduct a comprehensive inquiry of the current possibilities of using digital anthropometric patterns in image or video-based evidence.

A Virtual, 3D Multimodal Approach to Victim and Crime Scene Reconstruction

In the last two decades, forensic pathology and crime scene investigations have seen a rapid increase in examination tools due to the implementation of several imaging techniques, e.g., CT and MR scanning, surface scanning and photogrammetry. These tools encompass relatively simple visualization tools to powerful instruments for performing virtual 3D crime scene reconstructions. A multi-modality and multiscale approach to a crime scene, where 3D models of victims and the crime scene are combined, offers several advantages. A permanent documentation of all evidence in a single 3D environment can be used during the investigation phases (e.g., for testing hypotheses) or during the court procedures (e.g., to visualize the scene and the victim in a more intuitive manner). Advanced computational approaches to understand what might have happened during a crime can also be applied by, e.g., performing a virtual animation of the victim in the actual context, which can provide important information about possible dynamics during the event. Here, we present an overview of the different techniques and modalities used in forensic pathology in conjunction with crime scene investigations. Based on our experiences, the advantages and challenges of an image-based multi-modality approach will be discussed, including how their use may introduce new visualization modalities in court, e.g., virtual reality (VR) and 3D printing. Finally, considerations about future directions in research will be mentioned.

Forensic Anthropology and Archaeology in Denmark

In this paper, we provide a brief overview of the status of forensic anthropology and forensic archeology in Denmark, as well as related information about education, research, and skeletal collections. Forensic anthropologists mainly deal with the examination of unidentified skeletal remains. Some special tasks include cranial trauma analysis of the recently deceased, advanced 3D visualization from CT scanning of homicide cases, and stature estimation of perpetrators using surveillance videos. Forensic anthropologists are employed at one of Denmark’s three departments of forensic medicine (in Copenhagen, Odense, and Aarhus) and have access to advanced imaging equipment (e.g., CT and MR scanning, surface scanners, and 3D printers) for use in both their requisitioned work and their research. Extensive research is conducted on different topics, such as the health and diseases of past populations, age estimation, and human morphology. Research is based on skeletal material from the archeological collections housed in Copenhagen and Odense or on CT data from the recently deceased. There is no full degree in forensic anthropology in Denmark, but elective courses and lectures are offered to students at different levels and to people from different professional backgrounds.

Forensic archaeology is a relatively new field of expertise in Denmark, and relevant cases are rare, with only one or two cases per year. No forensic archeologists are officially employed in any of the departments of forensic medicine. Until recently, the Special Crime Unit of the police handled crime scene investigations involving excavations, but with the option of enlisting the help of outside specialists, such as archaeologists, anthropologists, and pathologists. An official excavation work group was established in 2015 under the lead of the Special Crime Unit of the police with the aim of refining the methods and procedures used in relevant criminal investigations. The group is represented by five police officers from the Special Crime Scene Unit, a police officer from the National Police Dog Training center, the two archaeologists from Moesgaard Museum, a forensic anthropologist from the Department of Forensic Medicine (University of Copenhagen), and a forensic pathologist from the Department of Forensic Medicine (University of Aarhus).

Online Boosting-Based Target Identification among Similar Appearance for Person-Following Robots

It is challenging for a mobile robot to follow a specific target person in a dynamic environment, comprising people wearing similar-colored clothes and having the same or similar height. This study describes a novel framework for a person identification model that identifies a target person by merging multiple features into a single joint feature online. The proposed framework exploits the deep learning output to extract four features for tracking the target person without prior knowledge making it generalizable and more robust. A modified intersection over union between the current frame and the last frame is proposed as a feature to distinguish people, in addition to color, height, and location. To improve the performance of target identification in a dynamic environment, an online boosting method was adapted by continuously updating the features in every frame. Through extensive real-life experiments, the effectiveness of the proposed method was demonstrated by showing experimental results that it outperformed the previous methods.

Human height estimation from highly distorted surveillance image

Video surveillance camera (VSC) is an important source of information during investigations especially if used as a tool for the extraction of verified and reliable forensic measurements. In this study, some aspects of human height extraction from VSC video frames are analyzed with the aim of identifying and mitigating error sources that can strongly affect the measurement. More specifically, those introduced by lens distortion are present in wide-field-of-view lens such as VSCs. A weak model, which is not able to properly describe and correct the lens distortion, could introduce systematic errors. This study focuses on the aspect of camera calibration to verify human height extraction by Amped FIVE software, which is adopted by the Forensic science laboratories of Carabinieri Force (RaCIS), Italy. A stable and reliable approach of camera calibration is needed since investigators have to deal with different cameras while inspecting the crime scene. The performance of the software in correcting distorted images is compared with a technique of single view self-calibration. Both approaches were applied to several frames acquired by a fish-eye camera and then measuring the height of five different people. Moreover, two actual cases, both characterized by common low-resolution and distorted images, were also analyzed. The height of four known persons was measured and used as reference value for validation. Results show no significant difference between the two calibration approaches working with fish-eye camera in test field, while evidence of differences was found in the measurement on the actual cases.

Improving 3D-3D facial registration methods: potential role of three-dimensional models in personal identification of the living

Personal identification of the living from video surveillance systems usually involves 2D images. However, the potentiality of three-dimensional facial models in gaining personal identification through 3D-3D comparison still needs to be verified. This study aims at testing the reliability of a protocol for 3D-3D registration of facial models, potentially useful for personal identification. Fifty male subjects aged between 18 and 45 years were randomly chosen from a database of 3D facial models acquired through stereophotogrammetry. For each subject, two acquisitions were available; the 3D models of faces were then registered onto other models belonging to the same and different individuals according to the least point-to-point distance on the entire facial surface, for a total of 50 matches and 50 mismatches. RMS value (root mean square) of point-to-point distance between the two models was then calculated through the VAM® software. Intra- and inter-observer errors were assessed through calculation of relative technical error of measurement (rTEM). Possible statistically significant differences between matches and mismatches were assessed through Mann–Whitney test (p < 0.05). Both for intra- and inter-observer repeatability rTEM was between 2.2 and 5.2%. Average RMS point-to-point distance was 0.50 ± 0.28 mm in matches, 2.62 ± 0.56 mm in mismatches (p < 0.01). An RMS threshold of 1.50 mm could distinguish matches and mismatches in 100% of cases. This study provides an improvement to existing 3D-3D superimposition methods and confirms the great advantages which may derive to personal identification of the living from 3D facial analysis.

A comparison of reverse projection and PhotoModeler for suspect height analysis

The purpose of this study was to compare the accuracy and precision between the Reverse Projection and PhotoModeler methods for suspect height analysis. Thirty analysts were assigned to measure the heights of three different suspects, one for each method, with the suspects having been recorded standing at three different distances in a scene. For Reverse Projection, the analysts were provided with height scales to place and video-record at the same positions their suspects stood in at the test scene, so that frames could be extracted from the video and overlaid onto frames of the suspects to measure height. For PhotoModeler, analysts calibrated frames of the suspects using 3D point cloud data obtained from a laser scan of the scene, so that measurements could be made in PhotoModeler software. Errors were calculated for the measurements and compared using the Mann-Whitney U-test and Kruskal-Wallis H-test, which indicated significant differences for errors between the two methods (p = 0.0025 and p = 0.008). Reverse Projection yielded a greater range of error and tended to have higher standard deviations than PhotoModeler, but the overall accuracy between the two methods was found to be comparable. The majority of absolute measurement errors for both methods were less than 2 cm.

The effects of camera resolution and distance on suspect height analysis using PhotoModeler

The purpose of this study was to examine how camera resolution and suspect-camera distance affect the accuracy and precision of suspect height estimations using PhotoModeler software. Sixteen individuals were measured and recorded standing at 15 pre-set distances on 7 security cameras, each with a different resolution setting. A height scale was used to measure each individual's height prior to recording and was also used as a reference height. Height estimates were taken in PhotoModeler by extracting video frames that were calibrated using 3D point model data obtained from a laser scan of the test site. Errors were calculated for the measurements and compared using the Kruskal-Wallis H-test, which indicated significant differences for errors among different resolutions and distances (p < 0.01). Interaction plots, however, demonstrated little difference in errors for most resolutions and positions. The accuracy and precision of height estimates began to decrease with resolutions under 960H and distances over 36.5 m.

A virtual reality method for digitally reconstructing traffic accidents from videos or still images

With an increase in the number of traffic accidents and enhanced attention to the rule of law, technical appraisement to reconstruct traffic accidents is attracting increasing attention. Accident videos are important aspects in identification; however, we cannot reconstruct an accident scene onsite using video for many reasons. In this paper, we introduce a computer-based virtual reality method that can digitally reconstruct a traffic accident. This method employs accident videos to perform a three-dimensional (3D) reconstruction of accident scenes. Using video screenshots, it constructs a model of humans and vehicles in 3D space to achieve the goal of dynamic restoration. The results indicate that this method has relatively high accuracy, requires little time and is easy to use. In this paper, we analyse the sources of errors for this method and summarize the application conditions.

Stature estimation from tibia percutaneous length: New equations derived from a Mediterranean population

Stature is a fundamental anthropometric character to trace the biological profile of a person. In some cases, when dismembered or mutilated bodies are discovered in a forensic context, it is essential to estimate stature from single districts of the body. Nevertheless, to date and worldwide, there are only few population-specific studies on stature estimation from leg length and none of them concerns modern populations in southern Europe. We attempted to fill this gap, focusing on the estimation of stature from the length of the tibia in a Mediterranean population (Italians). We carried out the current study on a sample of 374 Italian university students of both sexes (age range: 19.9-34.4). Both, actual stature and percutaneous length of tibia were measured and new equations were developed for stature estimation. We tested separate regression equations for each sex, as well as an equation for remains, whose sex is unknown. To assess their reliability, the equations were tested on a holdout sample of 30 individuals from the same population. Moreover, results of new specific linear regression equations were compared to others from the literature. We demonstrated that the newly proposed formulae (for males and combined sexes) and the ones by Olivier (for females) provided the most reliable estimations of stature for southern Europeans.

Height estimations based on eye measurements throughout a gait cycle

Anthropometric measurements (e.g. the height to the head, nose tip, eyes or shoulders) of a perpetrator based on video material may be used in criminal cases. However, several height measurements may be difficult to assess as the perpetrators may be disguised by clothes or headwear. The eye height (EH) measurement, on the other hand, is less prone to concealment. The purpose of the present study was to investigate: (1) how the eye height varies during the gait cycle, and (2) how the eye height changes with head position. The eyes were plotted manually in APAS for 16 test subjects during a complete gait cycle. The influence of head tilt on the EH was investigated in 20 healthy men. Markers were attached to the face and the subjects were instructed to stand relaxed, tilt their head to the right, to the left, forward and backward. The marker data for the right eye were used to calculate the EH. The respective deviation and SD from the relaxed standing EH and the EH in the Frankfurt plane, left tilted, right tilted, forward tilted and backward tilted, in addition to the corresponding head tilt angles were calculated. There was no correlation between the height of the subject and the maximum vertical displacement of the EH throughout the gait cycle nor between height of the subjects and the variation of the EH throughout the gait cycle. The average maximum vertical displacement for the test subject group was 4.76 cm (± 1.56 cm). The average EH was lower when the subjects were standing in the relaxed position than in the Frankfurt plane. The average EH was higher in the relaxed position than when the subjects tilted their heads, except when they tilted their heads backwards. The subjects had a slightly larger range of motion to the right than to the left, which was not significant. The results of this study provide a range for eye height estimates and may be readily implemented in forensic case work. It can be used as a reference in height estimates in cases with height measurements based on time of the gait cycle and based on the degree of head tilt from video material. Our data also provide descriptive statistics which may be helpful when comparing eye height measurements of a perpetrator with one or more suspects.

Metric and morphological assessment of facial features: a study on three European populations

Identification from video surveillance systems is becoming more and more frequent in the forensic practice. In this field, different techniques have been improved such as height estimation and gait analysis. However, the most natural approach for identifying a person in everyday life is based on facial characteristics. Scientifically, faces can be described using morphological and metric assessment of facial features. The morphological approach is largely affected by the subjective opinion of the observer, which can be mitigated by the application of descriptive atlases. In addition, this approach requires one to investigate which are the most common and rare facial characteristics in different populations. For the metric approach further studies are necessary in order to point out possible metric differences within and between different populations. The acquisition of statistically adequate population data may provide useful information for the reconstruction of biological profiles of unidentified individuals, particularly concerning ethnic affiliation, and possibly also for personal identification. This study presents the results of the morphological and metric assessment of the head and face of 900 male subjects between 20 and 31 years from Italy, Germany and Lithuania. The evaluation of the morphological traits was performed using the DMV atlas with 43 pre-defined facial characteristics. The frequencies of the types of facial features were calculated for each population in order to establish the rarest characteristics which may be used for the purpose of a biological profile and consequently for personal identification. Metric analysis performed in vivo included 24 absolute measurements and 24 indices of the head and face, including body height and body weight. The comparison of the frequencies of morphological facial features showed many similarities between the samples from Germany, Italy and Lithuania. However, several characteristics were rare or significantly more or less common in one population compared to the other two. On the other hand, all measurements and indices, except for labial width and intercanthal-mouth index showed significant differences between the three populations. As far as comparisons with other samples are concerned, the three European Caucasian samples differed from North American Caucasian, African and Asian groups as concerns the frequency of the morphological traits and the mean values of the metric analysis. The metric and morphological data collected from three European populations may be useful for forensic purposes in the construction of biological profiles and in screening for personal identification.

A new computer-assisted technique to aid personal identification

The paper describes a procedure aimed at identification from two-dimensional (2D) images (video-surveillance tapes, for example) by comparison with a three-dimensional (3D) facial model of a suspect. The application is intended to provide a tool which can help in analyzing compatibility or incompatibility between a criminal and a suspect's facial traits. The authors apply the concept of "geometrically compatible images". The idea is to use a scanner to reconstruct a 3D facial model of a suspect and to compare it to a frame extracted from the video-surveillance sequence which shows the face of the perpetrator. Repositioning and reorientation of the 3D model according to subject's face framed in the crime scene photo are manually accomplished, after automatic resizing. Repositioning and reorientation are performed in correspondence of anthropometric landmarks, distinctive for that person and detected both on the 2D face and on the 3D model. In this way, the superimposition between the original two-dimensional facial image and the three-dimensional one is obtained and a judgment is formulated by an expert on the basis of the fit between the anatomical facial districts of the two subjects. The procedure reduces the influence of face orientation and may be a useful tool in identification.