This presentation is a review of the basics involved in the interpretation of the crime scene based on facts derived from the physical and testimonial evidence obtained from the scene. This presentation will demonstrate the need to thoroughly document the scene to prove the interpretation. Part of this documentation is based on photography and crime scene sketches. While the methodology is simple and well demonstrated in this presentation this aspect is one of the tasks least completed by most law enforcement agencies.

Law enforcement is responsible for investigating crimes, identifying and arresting the suspects, and presenting evidence to a judge and jury in court. In order to objectively perform these duties, police need to gather accurate information and clearly explain the crime scene and physical evidence in a court of law. Part of this information includes the documentation of the incident. Documenting an incident has always been divided into three categories: notes, sketch, and photographs. This method of recording crime scenes has been the standard for years. The major drawback, however, is that the visual documents of sketches and photographs are two dimensional. This greatly restricts the actual visualization of the incident requiring a careful cross referencing of the details in order to understand it.

A kit has been developed for the preservation of vital forensic evidence on a suspect following a serious assault, murder or other offense where contamination may occur. This also includes the handling of firearms, explosives and/or drugs.

An inexpensive, rapid and sensitive method for the field pre-concentration and subsequent analysis of high explosive residues from solid debris samples and from aqueous samples has been evaluated using solid-phase microextraction. Explosives studied included nitrobenzene, 2-nitrotoluene, 3- nitrotoluene, 4-nitrotoluene, 2,6-dinitrotoluene, 1,3- dinitrobenzene, 2,4-dinitrotoluene, trinitrotoluene, 1,3,5- trinitrobenzene, 4-amino-2,6-dinitrotoluene, 2-amino-4,6- dinitrotolene and tetryl.

Computers and network technology have become inexpensive and powerful tools that can be applied to a wide range of criminal activity. Computers have changed the world's view of evidence because computers are used more and more as tools in committing `traditional crimes' such as embezzlements, thefts, extortion and murder. This paper will focus on reviewing the current state-of-the-art of the data recovery and evidence construction tools used in both the field and laboratory for prosection purposes.

Today, digital signature technologies are being used in many aspects of information security, including proof of identity, authentication, authorization, integrity and non- repudiation. The combination of Kim & Spafford's early research, and the application of digital signatures and SmartCards to the preservation of computer based evidence, is the basis of this paper. We will discuss the technical and operational weaknesses of current methods employed to protect the integrity of computer evidence, and recommend solutions for simplifying and enhancing the evidence preservation and collection process.

Reports that money in general circulation is contaminated with cocaine have resulted in contaminated money theories purporting that any person carrying currency could potentially initiate a drug dog alert. Field tests on dozens of different drug detector dogs with widely varying breeds, ages and training regimes show a consistent threshold level of 1 - 10 (mu) g of methyl benzoate spiked along with cocaine on U.S. currency or 0.1 - 1 ng/sec methyl benzoate diffusion required to initiate an alert. No other substance studied to data has initiated consistent responses by the drug dogs studied.

The scientific examination of physical evidence is well recognized as a critical element in conducting successful criminal investigations and prosecutions. The forensic science field is an ever changing discipline. With the arrival of DNA, new processing techniques for latent prints, portable lasers, and electro-static dust print lifters, and training of evidence technicians has become more important than ever. These scientific and technology breakthroughs have increased the possibility of collecting and analyzing physical evidence that was never possible before. The problem arises with the collection of physical evidence from the crime scene not from the analysis of the evidence. The need for specialized units in the processing of all crime scenes is imperative. These specialized units, called crime scene units, should be trained and equipped to handle all forms of crime scenes. The crime scenes units would have the capability to professionally evaluate and collect pertinent physical evidence from the crime scenes.

A novel identification technique called antibody profiling was examined as an alternative to DNA-based methods for matching crime scene evidence to a suspect. This technique provides results within 2 hours, is 1/100 the cost of DNA tests, and does not require skilled technicians or expensive equipment. A matrix of 422 blood samples were prepared to mimic typical crime scene conditions and provide validation for the technique. The effects of sample size, drying temperature, binary and ternary blood mixtures, adulteration with chemicals, and placement on a variety of surfaces were examined. Using the antibody profiling method, 91% of the 422 samples were correctly identified. In addition, binary blood mixtures could be identified with up to 40% contaminating blood. Temperatures at or above 60 degree(s)C and the presence of soil in the samples interfered with the ability to correctly identify samples. In this study, the antibody profiling technique was shown to be an excellent alternative to DNA-based identification methods. This method will find applications in situations where results are needed rapidly, where it is necessary to screen multiple suspects, and in remote areas where the equipment and technical skills needed for DNA testing are not available.

Imaging time-of-flight secondary ion mass spectrometry (SIMS) was used to evaluate the detection of contaminant chemicals on the surfaces of single synthetic textile and canine hair fibers. The results of the study showed that a variety of chemical classes can be detected. Both cocaine and heroin could be easily observed as intact protonated molecules ([M + H]⁺) in the cation spectra acquired from textile fibers. Two organophosphates were evaluated: malathion, which is a common pesticide, and pinacolyl methyl phosphonic acid (PMPA), which is the principal degradation product of the nerve agent soman (a close relative of sarin). Malathion could be observed as (CH₃O)₂P(equalsS)S^-, which is formed by thiophosphate cleavage of the intact malathion. PMPA is observed as the conjugate base ([PMPA - H]^-). Surfactant chemicals found in hair care products were successfully detected on single hair fibers. Specifically, alkyl sulfates, ethoxylated alkyl sulfates, silicones, and alkylammonium compounds could be readily identified in spectra acquired from single hair fiber samples exposed to shampoo and/or conditioner. Generally, the results of the study show that imaging SIMS is applicable to single fiber analysis, for a range of adsorbed compound types. The forensic application of this instrumental approach has not been widely recognized. However, the ability of the technique to acquire specific chemical information from trace samples clearly points to applications where the need for chemical analysis is great, but the amount of sample is limited.

A variety of paint and fingernail polish samples, which were visually similar, but had different chemical compositions, were analyzed using three static secondary ion mass spectrometry (SIMS) techniques. These techniques included: (1) high spatial resolution/high mass resolution imaging time-of-flight SIMS; (2) analysis of stabilized high mass secondary ions with an ion trap SIMS capable of MS/MS; (3) qualitative characterization using a quadrupole SIMS with `pulsed extraction' charge compensation. In some cases, distinguishing between different coatings was easily achieved because of the presence of dominant ions derived from the components of the coating materials in the SIMS spectra. In other instances, coating distinction was difficult within a product group because of spectral complexity; for this reason, multivariate statistical techniques were employed, which allowed meaningful classification of spectra. Partial Least Squares and Principle Component Analysis were applied to quadrupole SIMS data. When using Partial Least Squares analysis reasonably accurate coating identification was achieved with the preliminary data with overall correct identifications at greater than 90% sensitivity.

The Technical Working Group for Materials Analysis (TWGMAT) is supporting an international forensic automotive paint database. The Federal Bureau of Investigation and the Royal Canadian Mounted Police (RCMP) are collaborating on this effort through TWGMAT. This paper outlines the support and further development of the RCMP's Automotive Paint Database, `Paint Data Query'. This cooperative agreement augments and supports a current, validated, searchable, automotive paint database that is used to identify make(s), model(s), and year(s) of questioned paint samples in hit-and-run fatalities and other associated investigations involving automotive paint.

During the past five years work in our laboratory has been concentrated on developing SERRS spectroscopy and making it a simple and robust technique for the analyses of colorants. It has proved to be highly discriminative, extremely sensitive and possible to identify dyes in mixtures without their prior separation. Additionally, by using concentrated silver colloid solutions, in-situ analyses have now been accomplished with minimal or in some cases no visual destruction of the item being examined and with virtually no background interference from the surfaces on which the stains or smears have been deposited. To illustrate the methodology and the potential of SERRS various applications including the in-situ analyses of the dyes on cotton fibers and stains from cosmetics, shoe polishes, inks and drinks on various surfaces are presented.

Flat glass is a common type of evidence collected from the scenes of crimes such as burglaries, vandalism, and hit-and- run accidents. The usefulness of such evidence lies in the ability to associate the glass from the scene (or a suspect) to the original source. Physical and chemical analysis of the glass can be used for discrimination between the possible sources of glass. If the sample is large enough, physical attributes such as fracture matches, density, color, and thickness can be employed for comparison between a recovered fragment(s) to the suspect source. More commonly, refractive index (RI) comparisons are employed. Due to the improved control over glass manufacturing processes, RI values often cannot differentiate glasses where approximately 6 - 9% of casework samples are not expected to be distinguished by RI alone even if they originated from different sources. Employing methods such as NAA, XRF, ICP-AES, and ICP-MS for the comparison of trace elemental compositions has been shown to be more discriminating than RI comparisons. The multielement capability and the sensitivity of ICP-AES and ICP-MS provide for excellent discrimination power. In this work, the sources of variability in ICP-MS of glass analysis are investigated to determine possible sources of variation. The sources of variation examined include errors due to sample preparation, instrument accuracy and precision, and interlaboratory reproducibility. Other sources of variation include inhomogeneity across a sheet of glass from the same source. Analysis of variance has been applied to our ICP-MS analysis of NIST standards and to the interlaboratory comparisons of float glass samples collected across a sheet in a production facility. The results of these experiments allows for a more accurate interpretation of forensic glass data and a better understanding of the discriminating power (absolute and practical) of ICP-MS.

A technical overview of the Integrated Ballistic Identification System image acquisition hardware, image storage, case data input, `surface signature' analysis, and correlation scoring to an image database is described. Pilot evaluations by law enforcement, the impact of this technology in actual laboratory use, and the future use of a portable instrument in the field crime scene is also presented.

Several systems exist for collecting spent ammunition for forensic investigation. These databases store images of cartridge cases and the marks on them. The research in this paper is focused on the different methods of feature selection and pattern recognition that can be used for comparison. For automatic comparison of these images it is necessary to extract firstly the useful parts of the images. On databases of 3800 images several processing steps have been tested and compared. The results and methods, which have been implemented, are presented. The usual correlation methods based on gray values of all relevant image data have been tested. They were useful in the database. Further invariant image descriptors and the a trous wavelet transform have been implemented. These methods are promising, however more investigation is needed for the use of these methods.

The gunpowder residues that remain after the use of handguns or improvised explosive devices pose a challenge for the forensic investigator. Can these residues be reliably linked to a specific gunpowder or ammunition? We investigated the possibility by recovering and measuring the composition of organic additives in smokeless powder and its post-firing residues. By determining gunpowder additives such as nitroglycerin, dinitrotoluene, ethyl- and methylcentralite, and diphenylamine, we hope to identify the type of gunpowder in the residues and perhaps to provide evidence of a match to a sample of unfired powder. The gunpowder additives were extracted using an automated technique, pressurized fluid extraction (PFE). The conditions for the quantitative extraction of the additives using neat and solvent-modified supercritical carbon dioxide were investigated. All of the major gunpowder additives can be determined with baseline resolution using capillary electrophoresis (CE) with a micellar agent and UV absorbance detection. A study of candidate internal standards for use in the CE method is also presented. The PFE/CE technique is used to evaluate a new residue sampling protocol--asking shooters to blow their noses. In addition, an initial investigation of the compositional differences among unfired and post-fired .22 handgun residues is presented.

Capillary electrophoresis (CE) is a high resolution separation technique which is amenable to a wide variety of solutes, including compounds which are thermally degradable, non-volatile and highly polar, and is therefore well suited for drug analysis. Techniques which have been used in our laboratory include electrokinetic chromatography (ECC), free zone electrophoresis (CZE) and capillary electrochromatography (CEC). ECC, which uses a charged run buffer additive which migrates counter to osmotic flow, is excellent for many applications, including, drug screening and analyses of heroin, cocaine and methamphetamine samples. ECC approaches include the use of micelles and charged cyclodextrins, which allow for the separation of complex mixtures. Simultaneous separation of acidic, neutral and basic solutes and the resolution of optical isomers and positional isomers are possible. CZE has been used for the analysis of small ions (cations and anions) in heroin exhibits. For the ECC and CZE experiments performed in our laboratory, uncoated capillaries were used. In contrast, CEC uses capillaries packed with high performance liquid chromatography stationary phases, and offers both high peak capacities and unique selectivities. Applications include the analysis of cannabinoids and drug screening. Although CE suffers from limited concentration sensitivity, it is still applicable to trace analysis of drug samples, especially when using injection techniques such as stacking, or detection schemes such as laser induced fluorescence and extended pathlength UV.

In this paper, we present data comparing a variety of different conditions for extracting ignitable liquid residues from simulated fire debris samples in order to optimize the conditions for using Solid Phase Microextraction. A simulated accelerant mixture containing 30 components, including those from light petroleum distillates, medium petroleum distillates and heavy petroleum distillates were used to study the important variables controlling Solid Phase Microextraction (SPME) recoveries. SPME is an inexpensive, rapid and sensitive method for the analysis of volatile residues from the headspace over solid debris samples in a container or directly from aqueous samples followed by GC. The relative effects of controllable variables, including fiber chemistry, adsorption and desorption temperature, extraction time, and desorption time, have been optimized. The addition of water and ethanol to simulated debris samples in a can was shown to increase the sensitivity when using headspace SPME extraction. The relative enhancement of sensitivity has been compared as a function of the hydrocarbon chain length, sample temperature, time, and added ethanol concentrations. The technique has also been optimized to the extraction of accelerants directly from water added to the fire debris samples. The optimum adsorption time for the low molecular weight components was found to be approximately 25 minutes. The high molecular weight components were found at a higher concentration the longer the fiber was exposed to the headspace (up to 1 hr). The higher molecular weight components were also found in higher concentrations in the headspace when water and/or ethanol was added to the debris.

The use of a new Magnetic Media Imaging Instrument (MMII) as applied to the forensic investigation of audio tapes is reported. The MMII is an instrument that produces a 2D image of the magnetic fields on a segment of tape up to 5 cm long, with a resolution of about 3 microns. The dynamic range exceeds 40 dB. This allows the visualization of important characteristics such as stop marks with much greater detail than is possible with conventional means such as ferrofluids. Results of tests representing typical forensic examinations are presented.

In this paper, we describe the use and processing of image sequences for forensic investigations. For real-time digitization and storage of uncompressed movie files we use a computer equipped with a disk array system. The number of tape playbacks is minimized, while the use of still image/pause options of the VCR can be avoided completely.

Contrast enhancement methods have a long history of use in image processing for forensics and have been used to effect in the evaluation patterned injury of the skin. Most contrast enhancement methods, however, were developed for the evaluation of greyscale images and involve the manipulation of one dimension of data at a time. Contrast enhancement in a three- or more dimensional space poses challenges to the implementation of histogram equalization and similar algorithms. A number of approaches to dealing with this problem have been suggested, including performing operations on each channel independently or by various color `explosion' methods. Our laboratory has been investigating dispersion- and diffusion-based methods by modeling changes in color space as biological processes. In short, we model the migration and dispersion of points in color space as migration and differentiation. In this model, biological differentiation signals are used for segmentation in color space (color quantization) and chemoattractant and diffusion models are used for swarming and dispersal. The results of this method are compared with more traditional methods. Implementation issues are discussed. Extensions to the use of reaction-diffusion equations for color-space segmentation are discussed.

The aim of the `Image and Signal Processing' section at the INCC/NICC is to develop their own software tools in image processing. This paper gives an overview of the image processing environment that has been developed for about two years in the section. Above all, the main idea of FIPE (Forensic Image Processing Environment) is to build an user- friendly graphical interface especially adapted to forensic expert needs. The forensic expert's work is task oriented which means that the method used depends upon which information is needed from the image. Therefore, FIPE is presented as a menu with different functions which the user can pick up. Among the different tools, classical image enhancement methods, e.g. LUT enhancement, linear and non- linear filters, are implemented. Zoom and color image processing, e.g. color extraction, are also available. A regularization method using FFT is also implemented for out- of-focus deblurring. The environment is currently used on a daily basis in our section and is in constant development.

Examiners in the Special Photographic Unit of the Federal Bureau of Investigation Laboratory Division conduct examinations of questioned photographic evidence of all types, including surveillance imagery recorded on film and video tape. A primary type of examination includes side-by- side comparisons, in which unknown objects or people depicted in the questioned images are compared with known objects recovered from suspects or with photographs of suspects themselves. Most imagery received in the SPU for such comparisons originate from time-lapse video or film systems. In such circumstances, the delay between sequential images is so great that standard image summing and/or averaging techniques are useless as a means of improving image detail in questioned subjects or objects without also resorting to processing-intensive pattern reconstruction algorithms. Occasionally, however, the receipt of real-time video imagery will include a questioned object at rest. In such cases, it is possible to use relatively simple image averaging techniques as a means of reducing transient noise in the images, without further compromising the already-poor resolution inherent in most video surveillance images. This paper presents an example of one such case in which multiple images were averaged to reduce the transient noise to a sufficient degree to permit the positive identification of a vehicle based upon the presence of scrape marks and dents on the side of the vehicle.

Since late 1996 the Forensic Identification Services Section of the Ontario Provincial Police has been actively involved in state-of-the-art image capture and the processing of video images extracted from crime scene videos. The benefits and problems of this technology for video analysis are discussed. All analysis is being conducted on SUN Microsystems UNIX computers, networked to a digital disk recorder that is used for video capture. The primary advantage of this system over traditional frame grabber technology is reviewed. Examples from actual cases are presented and the successes and limitations of this approach are explored. Suggestions to companies implementing security technology plans for various organizations (banks, stores, restaurants, etc.) will be made. Future directions for this work and new technologies are also discussed.

To reliably perform photocomparisons of heads or skulls, it is important to position the head or skull corresponding to the posture of the head on the photograph. An explanation is given of the three-point analysis as a tool for positioning. This three-point analysis, taken from the head on the photograph, allows us to position heads or skulls in front of a camera corresponding to those points of the head on the photograph. A 24 X 36 mm camera with the three-point analysis within the viewfinder appeared to be suitable. This method was developed because the performance of photo comparisons on living persons in different surroundings, and the necessity for multiple comparisons per person demanded easier to handle equipment and a simplified positioning method. A novelty is bringing this three-point analysis into a computer, and using it together with a digitized one-shot movable 3D photograph of the head or skull of a person. Some experiences about working with such a 3D photograph will be discussed. The 3D technique used, was developed by scientists of the Roman Catholic University of Leuven, Belgium.

Person identification by using biometric methods based on image sequences, or still images, often requires a controllable and cooperative environment during the image capturing stage. In the forensic case the situation is more likely to be the opposite. In this work we propose a method that makes use of the anthropometry of the human body and human actions as cues for identification. Image sequences from surveillance systems are used, which can be seen as monocular image sequences. A 3D deformable wireframe body model is used as a platform to handle the non-rigid information of the 3D shape and 3D motion of the human body from the image sequence. A recursive method for estimating global motion and local shape variations is presented, using two recursive feedback systems.

In this paper, we present some results from a study in progress on methods for the measurement of the length of a robber in a surveillance video image. A calibration tool was constructed for the calibration of the camera. Standard procedures for computing the lens distortion, image projection parameters and the length of the robber have been implemented in Mathematika. These procedures are based on the use of pixel coordinates of markers on the calibration tool, the robber's head (and, optionally, his feet) and an estimation of his position in the coordinate system that is defined by the calibration tool.

We have studied a multi-frame image processing method for recognition of degraded car license plate images. In actual situations, various distortions such as geometric distortions or various blurs are imposed on recorded images. Therefore in order to apply the multi-frame image processing method to moving objects, it is necessary to correct geometric distortions. When images are blurred with out-of- focusing or motion blur, it is necessary to remove or suppress such distortions. In this report, we examined the geometric distortion correction and deblurring for multi- frame image processing and confirmed the effectiveness of the method. The example for real videotaped images is also shown in this report. As a result, it is recognized that the multi-frame image processing method is effective for recognizing characters on moving car license plates.

This paper describes a recent study conducted to investigate the reproducibility of toolmarks left by drill bits. This paper focuses on the automated analysis aspect of the study, and particularly the advantages of using decision fusion methods in the comparisons. To enable the study to encompass a large number of samples, existing technology was adapted to the task of automatically comparing the test impressions. Advanced forensic pattern recognition algorithms that had been developed for the comparison of ballistic evidence in the DRUGFIRE^TM system were modified for use in this test. The results of the decision fusion architecture closely matched those obtained by expert visual examination. The study, aided by the improved pattern recognition algorithm, showed that drill bit impressions do contain reproducible marks. In a blind test, the DRUGFIRE pattern recognition algorithm, enhanced with the decision fusion architecture, consistently identified the correct bit as the source of the test impressions.

The user of a software application does not care how the data are stored in memory, except perhaps if excess memory is used. What the user cares strongly about, however, is how the data are rendered on the screen and how the data can be located, inspected, analyzed and further processed. The majority of image processing software has adopted the use of layering for these purposes. This includes some kind of z- ordering for displaying images on top of each other and tools and techniques that lets the user access and manipulate the different layers. In this paper we will discuss the next step from simple layering: the concept of data object compositions. Finally we will see how to access some of the special classes' functionality. A lot of classes in our hierarchy will have some special abilities that aren't covered by the common interface. These abilities have to be exposed to the user.

The aim of this work is the removal of distracting background patterns from forensic evidence so that the evidence is rendered more visible. An example is the image of a finger print on a non-periodic background. The method involves registering the image with a control image of the background pattern that we seek to remove. A statistical comparison of the registered images identifies the latent mark.

In this paper a new measurement algorithm is presented which generates height measurements and their associated errors from a single known physical measurement in an image. The method draws on results from projective geometry and computer vision. A height measurement is obtained from each frame of the video. A `stereo like' correspondence between images is not required. Nor is any explicit camera calibration. The accuracy of the algorithm is demonstrated by a number of examples when ground truth is known. Finally, the height measurements and their variation are described for a person in motion. We draw attention to the uncertainty in heights associated with humans in motion, and the limitations of using this description for identification.