Discriminating natural images and computer generated graphics based on the impact of CFA interpolation on the correlation of PRNU

To discriminate natural images from computer generated graphics, a novel identification method based on the features of the impact of color filter array (CFA) interpolation on the local correlation of photo response non-uniformity noise (PRNU) is proposed. As CFA interpolation generally exists in the generation of natural images and it imposes influence on the local correlation of PRNU, the differences between the PRNU correlations of natural images and those of computer generated graphics are investigated. Nine dimensions of histogram features are extracted from the local variance histograms of PRNU to represent the identification features. The discrimination is accomplished by using a support vector machine (SVM) classifier. Experimental results and analysis show that it can achieve an average identification accuracy of 99.43%, and it is robust against scaling, JPEG compression, rotation and additive noise. Thus, it has great potential to be used in image source pipelines forensics.     

Computer‐aided Dental Identification: An Objective Method for Assessment of Radiographic Image Similarity*

Abstract: A pilot study evaluated a computer‐based method for comparing digital dental images, utilizing a registration algorithm to correct for variations in projection geometry between images prior to a subtraction analysis. A numerical assessment of similarity was generated for pairs of images. Using well‐controlled laboratory settings, the method was evaluated as to its ability to identify the correct specimen with positive results. A subsequent clinical study examined longitudinal radiographic examinations of selected anatomical areas on 47 patients, analyzing the computer‐based method in making the correct identification based upon a threshold level of similarity. The results showed that at a threshold of 0.855, there were two false negative and two false positive identifications out of 957 analyses. Based on these initial findings, 25 dental records having two sets of full mouth series of radiographs were selected. The radiographs were digitized and grouped into six anatomical regions. The more recent set of films served as postmortem images. Each postmortem image was analyzed against all other images within the region. Images were registered to correct for differences in projection geometry prior to analysis. An area of interest was selected to assess image similarity. Analysis of variance was used to determine that there was a significant difference between images from the same individual and those from different individuals. Results showed that the threshold level of concordance will vary with the anatomical region of the mouth examined. This method may provide the most objective and reliable method for postmortem dental identification using intra‐oral images.     

Digital imaging methods as an aid in dental identification of human remains

The physical comparison of known (K) and questioned (Q) evidence samples is an accepted tool in numerous forensic identification disciplines (1). A subset of this process is the use of antemortem and postmortem dental radiographs to identify unidentified human remains. This method has been generally accepted for decades (2). The outcome is performed with a considerable degree of accuracy, due in part to a finite pool of possible candidates for identification derived via the NCIC database, passenger lists, and law enforcement Missing Persons reports. This paper describes a dental identification comparison protocol that incorporated digital imaging technology in this process. The computer was used to create digital exemplars of the K and Q evidence that were spatially and quantitatively compared (3). The digital mode allowed direct metric and morphologic comparison through the aid of a digital camera, desktop computer, monitor, and printer. The well-known computer program Adobe Photoshop 5.0 (4) was used to process the digital information in two forensic cases described in this paper. It is a commercially available digital imaging editing program that is operated on laptop and desktop computers possessing sufficient chip speed and RAM (Pentium II or equivalent and at least 76MB RAM) to open the large-size files generated by high-resolution digital capture devices. This program accepts raster-based image formats (e.g. .JPG, .BMP). Photoshop is noted for its diverse imaging functions, which allow the computer monitor to be used as a comparison microscope when Q and K sample images are tiled side-by-side and/or superimposed. Two and three-dimensional Q and K evidence samples can be individually digitized and then independently resized to allow two-dimensional comparison. The investigator also has the ability to create magnified images (200% to 300%) when the original digital image has been captured at near photoquality resolution (300 dpi). The visual comparison of physical features on the computer monitor permits a large field of view and robust digital control over image quality. Photographic measurement and enhancement features of Adobe Photoshop mimics and in some circumstances surpasses the historic use of conventional photographic manipulation in forensic casework. This paper presents two cases processed via routine forensic odontology identification protocols. These protocols had minimal results due to limitations described in the case histories. The additional application of digital methods proved useful in the ultimate identification of these human remains.     

A Multimodal Fusion Approach for Bullet Identification Systems

In the field of forensic science, bullet identification is based on the fact that firing the cartridge from a barrel leaves exclusive microscopic striation on the fired bullets as the fingerprint of the firearm. The bullet identification methods are categorized in 2‐D and 3‐D based on their image acquisition techniques. In this study, we focus on 2‐D optical images using a multimodal technique and propose several distinct methods as its modalities. The proposed method uses a multimodal rule‐based linear weighted fusion approach which combines the semantic level decisions from different modalities with a linear technique that its optimized modalities weights have been identified by the genetic algorithm. The proposed approach was applied on a dataset, which includes 180 2‐D bullet images fired from 90 different AK‐47 barrels. The experimentations showed that our approach attained better results compared to common methods in the field of bullet identification.     

Quantitative Statistics and Identification of Tool‐Marks

This study was designed to establish a feature identification method of tool‐mark 2D data. A uniform local binary pattern histogram operator was developed to extract the tool‐mark features, and the random forest algorithm was adopted to identify these. The presented method was used to conduct five groups of experiments with a 2D dataset of known matched and nonmatched tool‐marks made by bolt clippers, cutting pliers, and screwdrivers. The experimental results show that the proposed method achieved a high rate of identification of the tool‐mark samples generated under identical conditions. The proposed method effectively overcomes the disadvantage of unstable illumination of 2D tool‐mark image data and avoids the difficulty in mark inspection caused by manually preset parameters in the existing methods, thus reducing the uncertainty of inspected results.     

Robust Shoeprint Retrieval Method Based on Local‐to‐Global Feature Matching for Real Crime Scenes

In this study, an automatic and robust crime scene shoeprint retrieval method is proposed. As most shoeprints left at crime scenes are randomly partial and noisy, crime scene shoeprint retrieval is a challenging task. To handle partial, noisy shoeprint images, we employ denoising deep belief network (DBN) to extract local features and use spatial pyramid matching (SPM) to obtain a local‐to‐global matching score. In this study, 536 query shoeprint images from crime scenes and a large scale database containing 34,768 shoeprint images are used to evaluate the retrieval performance. Experimental results show that the proposed method outperforms other state‐of‐the‐art methods in terms of retrieval accuracy, feature dimension, and retrieval speed. The proposed method achieves a cumulative match score (CMS) of 65.67% at top 10 which is 5.60% higher than the second best performing method.     

Identification Using Frontal Sinus by Three‐Dimensional Reconstruction from Computed Tomography*

Abstract: Radiographs and computed tomography (CT) images have been increasingly applied and accepted in forensic sciences. Along with radiographs and CT images, the frontal sinus is often used for the individual identification because of its unique and unchangeable characteristics. The purpose of this study is to define the usefulness of three‐dimensional (3‐D) images of the frontal sinus for identification. CT images from 119 Korean cadavers were built up for 3‐D reconstruction and surveyed with 15 measurements. The total volume of the sinus, some nonmetric characteristics, and the bilateral asymmetry index in men differed significantly from those in women (p < 0.05). The digit codes, six sections and 10‐digit number, were almost able to accurately identify individuals (98%). This study showed a statistical difference between the sexes and classified the fused and prominent middle sinuses for the first time. This proposed method for identification is more accurate than those used in other studies.     

The Possibilities and Limitations of Forensic Hand Comparison

On recordings of certain crimes, the face is not always shown. In such cases, hands can offer a solution, if they are completely visible. An important aspect of this study was to develop a method for hand comparison. The research method was based on the morphology, anthropometry, and biometry of hands. A new aspect of this study was that a manual and automated test were applied, which, respectively, assess many features and provide identification rates quickly. An important observation was that good quality images can provide sufficient hand details. The most distinctive features were the length/width ratio, the palm line pattern and the quantity of highly distinctive features present, and how they are distributed. The results indicate that experience did not improve the identification rates, while the manual test did. Intra‐observer variability did not influence the results, whereas hands of relatives were frequently misjudged. Both tests provided high identification rates.     

Digital camera identification using PRNU: A feature based approach

Source camera identification is one of the emerging field in digital image forensics, which aims at identifying the source camera used for capturing the given image. The technique uses photo response non-uniformity (PRNU) noise as a camera fingerprint, as it is found to be one of the unique characteristic which is capable of distinguishing the images even if they are captured from similar cameras. Most of the existing PRNU based approaches are very sensitive to the random noise components existing in the estimated PRNU, and also they are not robust when some simple manipulations are performed on the images. Hence a new feature based approach of PRNU is proposed for the source camera identification by choosing the features which are robust for image manipulations. The PRNU noise is extracted from the images using wavelet based denoising method and is represented by higher order wavelet statistics (HOWS), which are invariant features for image manipulations and geometric variations. The features are fed to support vector machine classifiers to identify the originating source camera for the given image and the results have been verified by performing ten-fold cross validation technique. The experiments have been carried out using the images captured from various cell phone cameras and it demonstrated that the proposed algorithm is capable of identifying the source camera of the given image with good accuracy. The developed technique can be used for differentiating the images, even if they are captured from similar cameras, which belongs to same make and model. The analysis have also showed that the proposed technique remains robust even if the images are subjected to simple manipulations or geometric variations.     

Identification of document paper using hybrid feature extraction

Document forgery is a significant issue in Korea, with around ten thousand cases reported every year. Analyzing paper plays a crucial role in examining questionable documents such as marketable securities and contracts, which can aid in solving criminal cases of document forgery. Paper analysis can also provide essential insights in other types of criminal cases, serving as an important clue for solving cases such as the source of a blackmail letter. The papermaking process generates distinct forming fabric marks and formations, which are critical features for paper classification. These characteristics are observable under transmitted light and are created by the forming fabric pattern and the distribution of pulp fibers, respectively. In this study, we propose a novel approach for paper identification based on hybrid features. This method combines texture features extracted from images converted using the gray-level co-occurrence matrix (GLCM) approach and a convolutional neural network (CNN), with another set of features extracted by the CNN using the same images as input. We applied the proposed method to classification tasks for seven major paper brands available in the Korean market, achieving an accuracy of 97.66%. The results confirm the applicability of this method for visually inspecting paper products and demonstrate its potential for assisting in solving criminal cases involving document forgery.     

Virtual Tool Mark Generation for Efficient Striation Analysis,

This study introduces a tool mark analysis approach based upon 3D scans of screwdriver tip and marked plate surfaces at the micrometer scale from an optical microscope. An open‐source 3D graphics software package is utilized to simulate the marking process as the projection of the tip's geometry in the direction of tool travel. The edge of this projection becomes a virtual tool mark that is compared to cross‐sections of the marked plate geometry using the statistical likelihood algorithm introduced by Chumbley et al. In a study with both sides of six screwdriver tips and 34 corresponding marks, the method distinguished known matches from known nonmatches with zero false‐positive matches and two false‐negative matches. For matches, it could predict the correct marking angle within ±5–10°. Individual comparisons could be made in seconds on a desktop computer, suggesting that the method could save time for examiners.     

Root Morphology and Anatomical Patterns in Forensic Dental Identification: A Comparison of Computer‐Aided Identification with Traditional Forensic Dental Identification*

Abstract: An online forensic dental identification exercise was conducted involving 24 antemortem–postmortem (AM–PM) dental radiograph pairs from actual forensic identification cases. Images had been digitally cropped to remove coronal tooth structure and dental restorations. Volunteer forensic odontologists were passively recruited to compare the AM–PM dental radiographs online and conclude identification status using the guidelines for identification from the American Board of Forensic Odontology. The mean accuracy rate for identification was 86.0% (standard deviation 9.2%). The same radiograph pairs were compared using a digital imaging software algorithm, which generated a normalized coefficient of similarity for each pair. Twenty of the radiograph pairs generated a mean accuracy of 85.0%. Four of the pairs could not be used to generate a coefficient of similarity. Receiver operator curve and area under the curve statistical analysis confirmed good discrimination abilities of both methods (online exercise = 0.978; UT‐ID index = 0.923) and Spearman’s rank correlation coefficient analysis (0.683) indicated good correlation between the results of both methods. Computer‐aided dental identification allows for an objective comparison of AM–PM radiographs and can be a useful tool to support a forensic dental identification conclusion.     

录音设备识别司法鉴定技术研究

目标随着手机和录音笔等数字录音设备的普及,数字录音已基本取代传统的模拟录音,成为录音司法鉴定主导性检材类型。数字录音资料作为视听资料的重要组成部分,其真实性司法鉴定新技术新方法的研究具有重要的理论意义和实际应用价值。方法研究基于录音设备识别的数字录音真实性鉴定技术,通过数字录音背景噪声片段的提取,计算录音设备相关的关键统计特征,包括采样直方图分布特征和平均频谱统计特征,并使用机器学习和模式分类方法对数字录音的载体即录音设备进行准确分类。结果实验中最高的分类准确性达到97.09%。在录音设备可分性研究成果基础上,提出应用于数字录音设备司法鉴定的可行实施方案。结论研究结果表明了基于信号统计特征分析的录音设备识别方法的可行性和准确性。     

Reliability of Craniofacial Superimposition Using Three‐Dimension Skull Model

Craniofacial superimposition is a technique potentially useful for the identification of unidentified human remains if a photo of the missing person is available. We have tested the reliability of the 2D‐3D computer‐aided nonautomatic superimposition techniques. Three‐dimension laser scans of five skulls and ten photographs were overlaid with an imaging software. The resulting superimpositions were evaluated using three methods: craniofacial landmarks, morphological features, and a combination of the two. A 3D model of each skull without its mandible was tested for superimposition; we also evaluated whether separating skulls by sex would increase correct identifications. Results show that the landmark method employing the entire skull is the more reliable one (5/5 correct identifications, 40% false positives [FP]), regardless of sex. However, the persistence of a high percentage of FP in all the methods evaluated indicates that these methods are unreliable for positive identification although the landmark‐only method could be useful for exclusion.     

Malingering by Proxy: A Literature Review and Current Perspectives

Malingering by proxy (MAL‐BP) is a form of maltreatment that involves a caregiver who fabricates or induces signs or symptoms in a child, dependent adult, or pet in pursuit of external, tangible incentives. Rarely studied, MAL‐BP has an unknown prevalence, and is a challenging diagnosis for healthcare professionals. Therefore, a comprehensive computer literature search and review was conducted. The review uncovered a total of sixteen case reports of MAL‐BP (eleven human, five veterinary). The motive for malingering was financial in all human cases and medication‐seeking in all veterinary cases. Although the strategies employed differed among the identified cases, common themes regarding the best approach to identification of MAL‐BP cases became evident. A comprehensive workup including a thorough history, physical examination, appropriate neuropsychological testing, and relevant collateral information forms the basis of an effective identification strategy. The optimal method of management is currently unclear due to a relative paucity of data and guidelines. However, management of these cases would likely include a team‐based approach with a prudent assessment of safety for the proxy and a low threshold for referral to appropriate services. Long‐term follow‐up is essential and should be approached from a biopsychosocial perspective. Attention, research, and guidance on this topic are needed to develop further evidence‐based guidelines for the identification and management of MAL‐BP.     

Photo‐Realistic Statistical Skull Morphotypes: New Exemplars for Ancestry and Sex Estimation in Forensic Anthropology

Graphic exemplars of cranial sex and ancestry are essential to forensic anthropology for standardizing casework, training analysts, and communicating group trends. To date, graphic exemplars have comprised hand‐drawn sketches, or photographs of individual specimens, which risks bias/subjectivity. Here, we performed quantitative analysis of photographic data to generate new photo‐realistic and objective exemplars of skull form. Standardized anterior and left lateral photographs of skulls for each sex were analyzed in the computer graphics program Psychomorph for the following groups: South African Blacks, South African Whites, American Blacks, American Whites, and Japanese. The average cranial form was calculated for each photographic view, before the color information for every individual was warped to the average form and combined to produce statistical averages. These mathematically derived exemplars—and their statistical exaggerations or extremes—retain the high‐resolution detail of the original photographic dataset, making them the ideal casework and training reference standards.     

A simple and effective image-statistics-based approach to detecting recaptured images from LCD screens

It is now extremely easy to recapture high-resolution and high-quality images from LCD (Liquid Crystal Display) screens. Recaptured image detection is an important digital forensic problem, as image recapture is often involved in the creation of a fake image in an attempt to increase its visual plausibility. State-of-the-art image recapture forensic methods make use of strong prior knowledge about the recapturing process and are based on either the combination of a group of ad-hoc features or a specific and somehow complicated dictionary learning procedure. By contrast, we propose a conceptually simple yet effective method for recaptured image detection which is built upon simple image statistics and a very loose assumption about the recapturing process. The adopted features are pixel-wise correlation coefficients in image differential domains. Experimental results on two large databases of high-resolution, high-quality recaptured images and comparisons with existing methods demonstrate the forensic accuracy and the computational efficiency of the proposed method.