Nurliza Abdullah

Can Computer Vision Techniques be Applied to Automated Forensic Examinations? A Study on Sex Identification from Human Skulls Using Head CT Scans

Sex determination from human skeletal remains is a challenging problem in forensic anthropology. The human skull has been regarded as the second best predictor of sex because it contains several sexually dimorphic traits. Previous studies have shown that morphological assessment and morphometric analysis can be used to assess sex variation from dried skulls. With the availability of CT scanners, the field has seen increasing computer aided techniques in assisting these traditional forensic examinations. However, they largely remain at the level of providing a digital interface for landmarking for morphometric analysis. A recent research has applied shape analysis techniques for morphological analysis on a specific part of the skull. In this paper, we endeavor to explore the application of computer vision techniques that have prominently been used in the field of 3D object recognition and retrieval, for providing alternative method to achieve sex identification from human skulls automatically. We suggest a possible framework for the whole process including multi-region representation of the skull with 3D shape descriptors, and particularly examined the role of 3D descriptors on sex identification accuracy. The experimental results on 100 head post mortem CT scans indicate the potential of 3D descriptors for skull sex classification. To the best of our knowledge, this is the first work to have approached skull sex prediction from this novel perspective.

Novel Group Variable Selection for Salient Skull Region Selection and Sex Determination

Sex determination in forensic analysis involves individual examination of different sites of the skull and combination of these sites to understand their impact on the estimation results. Conventionally, forensic experts perform a stepwise combination of several skull region assessment parameters to determine the most important regions with regard to the sex estimation results. This paper introduces a novel group variable selection algorithm: Graph Laplacian Based Group Lasso with split augmented Lagrangian shrinkage algorithm (SALSA) to automatically learn from data by structuring the data into a set of disjointed groups and imposing a number of group sparsity to discover the salient groups which influence the sex determination results. In order to attain this, the skull is partitioned into smaller regions (local regions) using fuzzy c-means (FCM), which are further arranged into clusters as structured groups. Then, we implement the SALSA based group lasso algorithm to impose sparsity on the groups. Our experiments are conducted on 100 skull samples obtained from hospital kuala lumpur (HKL) and the best estimation result obtained is 84.5%.

Stature estimation by CT scan images of metacarpal measurements in Malaysian Malay Males

Four basic biological profile were useful for human identification including stature. The aim is to develop regressions for stature using ten metacarpals in Malay males. The samples were retrieved from Hospital Kuala Lumpur from PMCT images, which comprised 200 males from 20 to 53 years old. The parameters were lengths often metacarpals measured by Osirix, in which strong correlations with stature were evident. Single linear regressions were developed with good accuracy, where both true and estimated statures showed close values by paired t-test. It is concluded that regression equations are useful to estimate stature in Malay males.

Quantification of spheno-occipital synchondrosis fusion in a contemporary Malaysian population

Timing of fusion of the spheno-occipital synchondrosis (SOS) is correlated with age. Previous research, however, has demonstrated variation in the timing of closure among different global populations. The present study aims to quantify the timing of SOS fusion in Malaysian individuals as visualised in multi-detector computed tomography (CT) scans and to thereafter formulate age estimation models based on fusion status. Anonymised cranial CT scans of 336 males and 164 females, aged 5-25 years, were acquired from the National Institute of Forensic Medicine, Hospital Kuala Lumpur and Department of Diagnostic Imaging, Hospital Sultanah Aminah. The scans were received in DICOM format and reconstructed into three-dimensional images using OsiriX. The SOS is scored as open, fusing endocranially, fusing ectocranially or completely fused. Statistical analyses are performed using IBM SPSS Statistics version 24. Transition analysis (Nphases2) is then utilised to calculate age ranges for each stage. To assess the reliability of an observation, intra- and inter-observer agreement is quantified using Fleiss Kappa and was found to be excellent (κ=0.785-0.907 and 0.812). The mean (SD) age for complete fusion is 20.84 (2.84) years in males and 19.78 (3.35) years in females. Transition ages between Stages 0 and 1, 1 and 2, and 2 and 3 in males are 12.52, 13.98 and 15.52 years, respectively (SD 1.37); in females, the corresponding data are 10.47, 12.26 and 13.80 years (SD 1.72). Complete fusion of the SOS was observed in all individuals above the age of 18 years. SOS fusion status provides upper and lower age boundaries for forensic age estimation in the Malaysian sample.

Manifestation of Coronary Atherosclerosis in Klang Valley, Malaysia: An Autopsy Study

Aims: The present study aimed to determine the epidemiological aspects of medico-legal autopsies and manifestation of coronary atherosclerosis. Methods: This was a cross sectional study involving 222 cases recruited from National Institute of Forensic Medicine (NIFM) Hospital Kuala Lumpur (HKL) and Department of Forensic Medicine Hospital Sungai Buloh (HSgB) for a period of 15 months, from December 2012 to April 2014. Sociodemographic and autopsy findings, including the cause and manner of death were documented. Results: Male and female subjects aged 18–70 years were recruited. Males contributed to 86% of the total subjects and comprised 61% of young adults. Road traffic accidents were the primary cause of death, contributing almost 50% of the subjects. One third of the cases comprised of death due to natural causes, wherein almost 75% of the subjects within this category succumbed to sudden cardiac death. Coronary artery disease (CAD) contributed to 60% of the sudden cardiac death (SCD). Single and double-vessel diseases were the most common pattern of atherosclerosis. In almost 80% of CAD cases, atherosclerosis affected the left anterior descending artery (LAD). Conclusion: Cardiovascular diseases were the most significant natural cause of sudden death with a staggering figure of 75%. CAD was the single most commonly encountered pathology within the SCD. Most cases presented with single and double-vessel diseases, observed in all subjects, as well as the young adult population.

Computer vision methods for cranial sex estimation

The objective of this study is to demonstrate through empirical evaluation the potential of a number of computer vision (CV) methods for sex determination from human skull. To achieve this, six local feature representations, two feature learnings, and three classification algorithms are rigorously combined and evaluated on skull regions derived from skull partitions. Furthermore, we introduce for the first time the application of multi-kernel learning (MKL) on multiple features for sex prediction from human skull. In comparison to the classical forensic methods, the results in this study are competitive, attesting to the suitability of CV methods for sex estimation. The proposed approach is fully automatic.

Stature Estimation from Lower Limb Anthropometry using Linear Regression Analysis: A Study on the Malaysian Population

INTRODUCTION Stature is one of the characteristics that could be used to identify human, besides age, sex and racial affiliation. This is useful when the body found is either dismembered, mutilated or even decomposed, and helps in narrowing down the missing persons identity. AIM The main aim of the present study was to construct regression functions for stature estimation by using lower limb bones in the Malaysian population. MATERIALS AND METHODS The sample comprised 87 adult individuals (81 males, 6 females) aged between 20 to 79 years. The parameters such as thigh length, lower leg length, leg length, foot length, foot height and foot breadth were measured. They were measured by a ruler and measuring tape. Statistical analysis involved independent t-test to analyse the difference between lower limbs in male and female. The Pearsons correlation test was used to analyse correlations between lower limb parameters and stature, and the linear regressions were used to form equations. The paired t-test was used to compare between actual stature and estimated stature by using the equations formed. RESULTS Using independent t-test, there was a significant difference (p< 0.05) in the measurement between males and females with regard to leg length, thigh length, lower leg length, foot length and foot breadth. The thigh length, leg length and foot length were observed to have strong correlations with stature with p= 0.75, p= 0.81 and p= 0.69, respectively. Linear regressions were formulated for stature estimation. Paired t-test showed no significant difference between actual stature and estimated stature. CONCLUSIONS It is concluded that regression functions can be used to estimate stature to identify skeletal remains in the Malaysia population.