A variable resolution video machine vision system has been built which will determine the 3-D shape of a firing pin from the dent it has made in the primer of a shell. The system can also compare primer dents in different shells and determine if a 3-D match exists. A real time optical 3-D shape recognition feature allows the operator to align the shells, determine if there is a match, and quantify the quality of the 3-D match. The system uses a laser interferometer to project variable spatial frequency structural illumination on the shells which are viewed by video cameras. A computer determines the 3-D shape of the primer dent and predicts the shape of the firing pin. Optical shape recognition and alignment is achieved by mixing the video data with a computer generated mask and using this signal in a spatial light modulator to diffract a laser beam onto a detector. The magnitude and shape of the detector output indicates in real time the quality of the match or can be used to align the shells. A demonstration of the validity of the 3-D shape prediction has been carried out with large scale firing pin and shell-primer models, and the shape of a real firing pin has been determined from an actual fired shell. The result is a mixed video-optical processing system that could be used for real-time shell and firing pin matching in the forensic laboratory. Limitations and improvements in the current system are discussed.
An instrument to produce images of the magnetic fields on selected areas of magnetic media is being developed. Although most types of planar media can be imaged, the immediate application is for magnetic tape. Here, it replaces the use of ferrofluids, giving much higher resolution and contrast. The resolution is 1 micron, with a dynamic range of 40 dB. The current instrument can examine an area of 5 cm multiplied by 1 cm, but larger areas could be handled with minor changes. Detailed images of most features of interest, such as stop marks on magnetic tape, can be produced in less than 15 minutes. Lower resolution images of approximately 5 cm lengths of tape for feature identification can be made in significantly less time. Details of the instrument's performance and example images of interest to the forensic community are presented.
A rapid, readily automated method has been developed for the quantitative analysis of phenobarbital from human liver tissues based on supercritical carbon dioxide extraction followed by fluorescence enzyme immunoassay. The method developed significantly reduces sample handling and utilizes the entire liver homogenate. The current method yields comparable recoveries and precision and does not require the use of an internal standard, although traditional GC/MS confirmation can still be performed on sample extracts. Additionally, the proposed method uses non-toxic, inexpensive carbon dioxide, thus eliminating the use of halogenated organic solvents.
A statistical approach to the significance of glass evidence is proposed using independent physicochemical measurements and chemometrics. Traditional interpretation of the significance of trace evidence matches or exclusions relies on qualitative descriptors such as 'indistinguishable from,' 'consistent with,' 'similar to' etc. By performing physical and chemical measurements with are independent of one another, the significance of object exclusions or matches can be evaluated statistically. One of the problems with this approach is that the human brain is excellent at recognizing and classifying patterns and shapes but performs less well when that object is represented by a numerical list of attributes. Chemometrics can be employed to group similar objects using clustering algorithms and provide statistical significance in a quantitative manner. This approach is enhanced when population databases exist or can be created and the data in question can be evaluated given these databases. Since the selection of the variables used and their pre-processing can greatly influence the outcome, several different methods could be employed in order to obtain a more complete picture of the information contained in the data. Presently, we report on the analysis of glass samples using refractive index measurements and the quantitative analysis of the concentrations of the metals: Mg, Al, Ca, Fe, Mn, Ba, Sr, Ti and Zr. The extension of this general approach to fiber and paint comparisons also is discussed. This statistical approach should not replace the current interpretative approaches to trace evidence matches or exclusions but rather yields an additional quantitative measure. The lack of sufficient general population databases containing the needed physicochemical measurements and the potential for confusion arising from statistical analysis currently hamper this approach and ways of overcoming these obstacles are presented.
Digital imaging is emerging as a powerful technology in law enforcement and forensic science. PhotoCDTM offers the opportunity to bring traditional photographic images to the computer as well as television. This provides an inexpensive and rapid method of effectively enlarging and locally correcting exposure for all of the photographs taken in an investigation. It also provides a relatively inexpensive method for providing all of the photographic evidence in an investigation to anyone who has a need for it.
Forensic video image analysis is a new scientific tool for perpetrator enhancement and identification in poorly recorded crime scene situations. Forensic video image analysis is emerging technology for law enforcement, industrial security and surveillance addressing the following problems often found in these poor quality video recorded incidences.
Computer aided drafting (CAD) programs have great potential for helping the forensic scientist. One of their most direct and useful applications is crime scene documentation, as an aid in rendering neat, unambiguous line drawings of crime scenes. Once the data has been entered, it can easily be displayed, printed, or plotted in a variety of formats. Final renditions from this initial data entry can take multiple forms and can have multiple uses. As a demonstrative aid, a CAD program can produce two dimensional (2-D) drawings of the scene from one's notes to scale. These 2-D renditions are court display quality and help to make the forensic scientists's testimony easily understood. Another use for CAD is as an analytical tool for scene reconstruction. More than just a drawing aid, CAD can generate useful information from the data input. It can help reconstruct bullet paths or locations of furniture in a room when it is critical to the reconstruction. Data entry at the scene, on a notebook computer, can assist in framing and answering questions so that the forensic scientist can test hypotheses while actively documenting the scene. Further, three dimensional (3-D) renditions of items can be viewed from many 'locations' by using the program to rotate the object and the observers' viewpoint.
The use of canines to signal money associated with drug trafficking is a major point of contention among dog handlers, forensic scientists, and the legal community particularly in light of recent reports that a significant amount of money in circulation is contaminated with detectable amounts of cocaine. The questions raised include: What exactly are the dogs alerting to? How selective are the dogs? How sensitive are the dogs? How reliable are the dogs? Tests with various volatile cocaine by-products confirm that the dominant chemical in cocaine odor is methyl benzoate. Field tests on fifteen different drug detector dogs with varying breeds, ages and training regimes show a consistent threshold level of 1 (mu) g of methyl benzoate spiked along with cocaine on U.S. currency required to initiate an alert. The majority of the canines did not alert to pharmaceutical grade cocaine even at levels as high as 1 g. Methyl benzoate is shown to evaporate rapidly from individual bills and is a function of the available surface area for wrapped currency. The canines tested were remarkably selective and reliable even under varying test conditions and using different delivery devices.
A major problem hindering criminal investigations is the lack of appropriate tools for proper crime scene investigations. Often locating important pieces of evidence means relying on the ability of trained detection canines. Development of analytical technology to uncover and analyze evidence, potentially at the scene, could serve to expedite criminal investigations, searches, and court proceedings. To address this problem, a new technology based on gas sensor arrays was investigated for its applicability to forensic and law enforcement problems. The technology employs an array of sensors that respond to volatile chemical components yielding a characteristic 'fingerprint' pattern representative of the vapor-phase composition of a sample. Sample aromas can be analyzed and identified using artificial neural networks that are trained on known aroma patterns. Several candidate applications based on known technological needs of the forensic and law enforcement communities have been investigated. These applications have included the detection of aromas emanating from cadavers to aid in determining time since death, drug detection for deterring the manufacture, sale, and use of drugs of abuse, and the analysis of fire debris for accelerant identification. The result to date for these applications have been extremely promising and demonstrate the potential applicability of this technology for forensic use.
Through a joint agreement with the Indiana-Marion County Forensic Laboratory Services Agency, the Institute for Forensic Imaging conducted a pilot program to investigate crime lab applications of a digital darkroom. IFI installed and staffed a state-of-the-art digital darkroom in the photography laboratory of the Indianapolis-Marion County crime lab located at Indianapolis, Indiana. The darkroom consisted of several high resolution color digital cameras, image processing computer, dye sublimation continuous tone digital printers, and CD-ROM writer. This paper describes the use of the digital darkroom in several crime lab investigations conducted during the program.
The U.S. Secret Service has been doing and supporting research in several areas of fingerprint visualization. The following is discussed: (1) developing ninhydrin analogues for visualizing latent prints on porous surfaces such as paper (with Dr. Madeleine Joullie, University of Pennsylvania); (2) exploring reflective UV imaging techniques as a no-treatment-required method for visualizing latent prints; (3) optimizing 'gun bluing' methods for developing latent prints on metal surfaces (such as spent cartridges); (4) investigating aqueous metal deposition methods for visualizing latent prints on multiple types of surfaces; and (5) studying methods of transferring latent print residues onto membranes.
The observation that the fingerprints of children disappear from surfaces more quickly than those of adults initiated a study to characterize the chemical components in fingerprints. Samples were obtained from about 50 individuals ranging in age from three to 64 by extracting chemicals from the fintertips using rubbing alcohol. Using combined gas chromatography/mass spectrometry, a wide range of compounds were identified. It was found that the chemical compositions of fingerprints were quite different in children and adults. In general, the samples obtained from children contained higher levels of relatively volatile free fatty acids. Samples from adults were found to have higher concentrations of less volatile long chain esters of fatty acids. These esters are thought to originate from sebaceous glands located on the face and the levels of these compounds increase substantially after puberty. In addition to these compounds, a variety of other compounds were observed that could be used to develop improved methods for fingerprint detection at a crime scene. Further, the observation of specific compounds raises the possibility of being able to identify personal traits (gender, habits, diseases, etc.) via the analysis of components in fingerprints and/or skin.
It is our aim to develop a universally applicable latent fingerprint detection method using lanthanide (rare-earth) complexes as a source of luminescence. Use of these lanthanide complexes offers advantages on several fronts, including benefits from large Stokes shifts, long luminescence lifetimes, narrow emissions, ability of sequential assembly of complexes, and chemical variability of the ligands. Proper exploitation of these advantages would lead to a latent fingerprint detection method superior to any currently available. These same characteristics also lend themselves to many of the problems associated with DNA processing in the forensic science context.
The visitor to any modern forensic science laboratory is confronted with equipment and processes that did not exist even 10 years ago: thermocyclers to allow genetic typing of nanogram amounts of DNA isolated from a few spermatozoa; scanning electron microscopes that can nearly automatically detect submicrometer sized particles of molten lead, barium and antimony produced by the discharge of a firearm and deposited on the hands of the shooter; and computers that can compare an image of a latent fingerprint with millions of fingerprints stored in the computer memory. Analysis of populations of physical evidence has permitted statistically minded forensic scientists to use Bayesian inference to draw conclusions based on a priori assumptions which are often poorly understood, irrelevant, or misleading. National commissions who are studying quality control in DNA analysis propose that people with barely relevant graduate degrees and little forensic science experience be placed in charge of forensic DNA laboratories. It is undeniable that high- tech has reversed some miscarriages of justice by establishing the innocence of a number of people who were imprisoned for years for crimes that they did not commit. However, this papers deals with the dangers of technology in criminal investigations.
The Internet is a very powerful and inexpensive tool that was created for the free distribution of knowledge and information. The Internet is a learning tool, a research tool, a virtual library without borders and membership requirements, a journal with instant publication, a help desk, and a newspaper/journal with current information. Very soon, when live audio and video transmission is perfected, the Internet also will be a live classroom and everyday conference. Forensic scientists, laboratories and colleges should make use of information already available on the Internet. They also should actively participate and contribute. Very few forensic scientists and laboratories have made their presence felt by setting up their home pages/web pages. But, there is tremendous growth during the past year. Immense benefits from Internet to forensic community are discussed along with the author's personal experience. Creating on-line searchable data bases in all specialties of forensic science is an urgent need. Leading forensic journals should take a lead and create on-line searchable indexes with abstracts. On line electronic publishing, collaborative research/paper publishing or editing is easy, fast, economical and convenient through the use of the Internet. Creation of Internet repositories of unpublished papers is an idea worth looking into. Internet also can be used to give training, re-training or advanced training to students/forensic scientists.
Graphic Data Systems Corporation (GDS Corp.) and Intellignet Graphics Solutions, Inc. (IGS) combined talents in 1995 to design and develop a MicroGDSTM application to support field investiations of crime scenes, such as homoicides, bombings, and arsons. IGS and GDS Corp. prepared design documents under the guidance of federal, state, and local crime scene reconstruction experts and with information from the FBI's evidence response team field book. The application was then developed to encompass the key components of crime scene investigaton: staff assigned to the incident, tasks occuring at the scene, visits to the scene location, photogrpahs taken of the crime scene, related documents, involved persons, catalogued evidence, and two- or three- dimensional crime scene reconstruction. Crime scene investigation, reporting, and reconstruction (CSIRR$CPY) provides investigators with a single applicaiton for both capturing all tabular data about the crime scene and quickly renderng a sketch of the scene. Tabular data is captured through ituitive database forms, while MicroGDSTM has been modified to readily allow non-CAD users to sketch the scene.
The National Law Enforcement and Corrections Technology Centers (NLECTC) support public law enforcement agencies with technology development, evaluation, planning, architecture, and implementation. The NLECTC Western Region has a particular emphasis on surveillance and imaging issues. Among its activities, working with government and industry, NLECTC-WR produces 'Guides to Best Practices and Acquisition Methodologies' that facilitate government organizations in making better informed purchasing and operational decisions. This presentation includes specific examples from current activities. Through these systematic procedures, it is possible to design solutions optimally matched to the desired outcomes and provide a process for continuous improvement and greater public awareness of success.
This presentation describes the use of portable gas chromatography/mass spectrometry for on-scene investigation and processing of clandestine laboratories. Clandestine laboratory investigations present special problems to forensic investigators. These crime scenes contain many chemical hazards that must be detected, identified and collected as evidence. Gas chromatography/mass spectrometry performed on-scene with a rugged, portable unit is capable of analyzing a variety of matrices for drugs and chemicals used in the manufacture of illicit drugs, such as methamphetamine. Technologies used to detect various materials at a scene have particular applications but do not address the wide range of samples, chemicals, matrices and mixtures that exist in clan labs. Typical analyses performed by GC/MS are for the purpose of positively establishing the identity of starting materials, chemicals and end-product collected from clandestine laboratories. Concerns for the public and investigator safety and the environment are also important factors for rapid on-scene data generation. Here is described the implementation of a portable multiple-inlet GC/MS system designed for rapid deployment to a scene to perform forensic investigations of clandestine drug manufacturing laboratories. GC/MS has long been held as the 'gold standard' in performing forensic chemical analyses. With the capability of GC/MS to separate and produce a 'chemical fingerprint' of compounds, it is utilized as an essential technique for detecting and positively identifying chemical evidence. Rapid and conclusive on-scene analysis of evidence will assist the forensic investigators in collecting only pertinent evidence thereby reducing the amount of evidence to be transported, reducing chain of custody concerns, reducing costs and hazards, maintaining sample integrity and speeding the completion of the investigative process.