This PDF file contains the front matter associated with SPIE Proceedings Volume 6741, including the Title Page, Copyright information, Table of Contents, Introduction (if any), and the Conference Committee listing.

Operators are key components in a Closed Circuit Television (CCTV) system, being the link between the system technology and its effective use. Operators' performance will largely determine the level of service provided by the system. There have been few studies testing operator performance, while much work has been done to test the performance of the technology. Previous work on CCTV operator performance carried out by the Home Office Scientific Development Branch (HOSDB) has used filmed video and subjects who knew they were undergoing testing, meaning subjects are likely to be concentrating harder on performing well. HOSDB believes that there is a need for a test that would be able to be routinely used in a CCTV control room throughout the course of a normal shift to provide management with operational performance data. Threat Image Projection (TIP) is routinely used in X-Ray baggage scanners to keep operators alert to possible threats. At random intervals, a threat target image is superimposed over the image of the baggage being screened. The operator then responds to this threat. A similar system could be used for CCTV operators. A threat image would be randomly superimposed over the live CCTV feed and the operator would be expected to respond to this.

There are large amount of CCTV cameras collecting colossal amounts of video data about people and their behaviour. However, this overwhelming amount of data also causes overflow of information if their content is not analysed in a wider context to provide selective focus and automated alert triggering. To date, truly semantics based video analytic systems do not exist. There is an urgent need for the development of automated systems to monitor holistically the behaviours of people, vehicles and the whereabout of objects of interest in public space. In this work, we highlight the challenges and recent progress towards building computer vision systems for holistic video detection in a distributed network of multiple cameras based on object localisation, categorisation and tagging from different views in highly cluttered scenes.

Civilian soft targets such as transportation systems are being targeted by terrorists using IEDs and suicide bombers. Having the capability to remotely detect explosives, precursors and other chemicals would enable these assets to be protected with minimal interruption of the flow of commerce. Mid-IR laser technology offers the potential to detect explosives and other chemicals in real-time and from a safe standoff distance. While many of these agents possess "fingerprint" signatures in the mid-IR (i.e. in the 3-20 micron regime), their effective interrogation by a practical, field-deployable system has been limited by size, complexity, reliability and cost constraints of the base laser technology. Daylight Solutions has addressed these shortcomings by developing compact, portable, broadly tunable mid-IR laser sources based upon external-cavity quantum cascade technology. This technology is now being applied by Daylight in system level architectures for standoff and remote detection of explosives, precursors and chemical agents. Several of these architectures and predicted levels of performance will be presented.

The Imagery Library for Intelligent Detection Systems (i-LIDS) is the UK government's new standard for Video Based Detection Systems (VBDS). The standard was launched in November 2006 and evaluations against it began in July 2007. With the first four i-LIDS scenarios completed, the Home Office Scientific development Branch (HOSDB) are looking toward the future of intelligent vision in the security surveillance market by adding a fifth scenario to the standard. The fifth i-LIDS scenario will concentrate on the development, testing and evaluation of systems for the tracking of people across multiple cameras. HOSDB and the Centre for the Protection of National Infrastructure (CPNI) identified a requirement to track targets across a network of CCTV cameras using both live and post event imagery. The Detection and Vision Systems group at HOSDB were asked to determine the current state of the market and develop an in-depth Operational Requirement (OR) based on government end user requirements. Using this OR the i-LIDS team will develop a full i-LIDS scenario to aid the machine vision community in its development of multi-camera tracking systems. By defining a requirement for multi-camera tracking and building this into the i-LIDS standard the UK government will provide a widely available tool that developers can use to help them turn theory and conceptual demonstrators into front line application. This paper will briefly describe the i-LIDS project and then detail the work conducted in building the new tracking aspect of the standard.

In this paper we present an approach for tracking people across non overlapping cameras. The approach proposed is based a multi-dimensional feature vector and its covariance, defining an appearance model of every detected blob in the network of cameras. The model integrates relative position, color and texture descriptors of each detected object. Association of objects across non-overlapping cameras is performed by matching detected objects appearance with past observations. Availability of tracking within every camera can further improve the accuracy of such association by matching several targets appearance models with detected regions. For this purpose we present an automatic clustering technique allowing to build a multi-valued appearance model from a collection of covariance matrices. The proposed approach does not require geometric or colorimetric calibration of the cameras. We will illustrate the method for tracking people in relatively crowded scenes in a collection of indoors cameras taken in a mass transportation site. We will present success and challenges yet to be addressed by the proposed approach.

This paper describes an automated video analysis based monitoring system with processing at the sensor edge to watch and report certain predetermined events and unusual activities in remote areas that may be in unfriendly zones. The prototype system developed here involves content extraction from video streams collected by unattended ground cameras, tracking of objects, detection of events, and assessment of scenes for anomalous situations. The application requirements impose efficiency constraints on video analysis algorithms due to low-power on the sensor processing board. We present efficient video analysis algorithms for detection, tracking and classification of objects, analysis of extracted object and scene information to detect specific events as well as anomalous or novel situations at the video camera level. Our multi-tier and modular video analysis approach uses fast a space-based peripheral vision component for quick spatially based tracking of objects, detailed object- or scene-based feature extractors and data driven Support Vector Machine (SVM) classifiers that handle feature-based analysis at multiple data levels. Our algorithms are developed and tested on PC platform but designed to match the processing and power limitations of the target hardware platform. The video object detection and tracking components have been implemented on Texas Instruments DM642 evaluation board for assessing the feasibility of the prototype system.

Establishing benchmark datasets, performance metrics and baseline algorithms have considerable research significance in gauging the progress in any application domain. These primarily allow both users and developers to compare the performance of various algorithms on a common platform. In our earlier works, we focused on developing performance metrics and establishing a substantial dataset with ground truth for object detection and tracking tasks (text and face) in two video domains -- broadcast news and meetings. In this paper, we present the results of a face detection and tracking algorithm on broadcast news videos with the objective of establishing a baseline performance for this task-domain pair. The detection algorithm uses a statistical approach that was originally developed by Viola and Jones and later extended by Lienhart. The algorithm uses a feature set that is Haar-like and a cascade of boosted decision tree classifiers as a statistical model. In this work, we used the Intel Open Source Computer Vision Library (OpenCV) implementation of the Haar face detection algorithm. The optimal values for the tunable parameters of this implementation were found through an experimental design strategy commonly used in statistical analyses of industrial processes. Tracking was accomplished as continuous detection with the detected objects in two frames mapped using a greedy algorithm based on the distances between the centroids of bounding boxes. Results on the evaluation set containing 50 sequences (≈ 2.5 mins.) using the developed performance metrics show good performance of the algorithm reflecting the state-of-the-art which makes it an appropriate choice as the baseline algorithm for the problem.

The method of object recognition is described by the example of objects of amplitude transparent type. The method is to obtain a photoanisotropic copy of object images on the polarization-sensitive material. At consequent illumination of the photoanisotropic copy with a parallel circularly polarized beam of nonactinic light, the transmitted light becomes elliptically polarized. It is shown that the characteristics of the summary polarization ellipse in the Fraunhofer diffraction region uniquely identify the given object. The real time determination of the characteristics of the summary polarization ellipse is made by means of diffraction gratings of anisotropic profile and by comparison of these characteristics with etalon from database.

Today, a large number of videos are collected using aerial platforms. These videos are used for various applications from agriculture surveys to disaster response, from surveillance and security to intelligence gathering. As the amount of aerial video increases, there is a need for systematic exploitation and effective management of the large aerial videos database. In this paper, we will introduce VideoQuest, an advanced aerial video exploitation and management system that provides real-time aerial video enhancement, archiving, indexing and analysis capabilities such as sensor metadata enhancement, moving target detection and tracking and event detection. To effectively and efficiently utilize archived aerial videos, VideoQuest also provides spatial, temporal and content based indexing. To quickly retrieve videos in a large-scale video database, the system summarizes aerial video hierarchically and based on content, such as objects, tracks and events extracted from videos. Additionally, VideoQuest allows user to interactively search and browse A large aerial video database through a "virtual fly control" GUI that dynamically assembles visual information according to user's needs. Using the VideoQeust system, a user can search and retrieve mission-relevant information several magnitudes faster than without using our system.

Latent prints on substrates of varying porosity are visualized by a host of physical, optical, and chemical techniques that are often used sequentially. Optical techniques include viewing the scattering of UV radiation by latent print residue on smooth surfaces and inducing natural (inherent) fluorescence of such residue; physical techniques include visualizing the latent print using diverse powders and vapors that selectively polymerize on the residue; and chemical techniques include reacting components of latent print residue with reagents that render color and/or fluorescence to the residue. Recent developments in visualizing techniques include the use of nanoparticles, hyperspectral imaging, and the development of more sensitive reagents for visualizing the water-soluble components of latent print residue (e.g., amino acids) and the water-insoluble components (e.g., lipids, dried proteins, etc.). The chemistry of latent print residue and current research in the methods to visualize its components will be discussed.

Primarily focused at military and security environments where there is a need to identify humans covertly and remotely; this paper outlines how recovering human gait biometrics from a multi-spectral imaging system can overcome the failings of traditional biometrics to fulfil those needs. With the intention of aiding single camera human gait recognition, an algorithm was developed to accurately segment a walking human from multi-spectral imagery. 16-band imagery from the image replicating imaging spectrometer (IRIS) camera system is used to overcome some of the common problems associated with standard change detection techniques. Fusing the concepts of scene segmentation by spectral characterisation and background subtraction by image differencing gives a uniquely robust approach. This paper presents the results of real trials with human subjects and a prototype IRIS camera system, and compares performance to typical broadband camera systems.

Due to a large increase in the video surveillance data recently in an effort to maintain high security at public places, we need more robust systems to analyze this data and make tasks like face recognition a realistic possibility in challenging environments. In this paper we explore a watch-list scenario where we use an appearance based model to classify query faces from low resolution videos into either a watch-list or a non-watch-list face. We then use our simple yet a powerful face recognition system to recognize the faces classified as watch-list faces. Where the watch-list includes those people that we are interested in recognizing. Our system uses simple feature machine algorithms from our previous work to match video faces against still images. To test our approach, we match video faces against a large database of still images obtained from a previous work in the field from Yahoo News over a period of time. We do this matching in an efficient manner to come up with a faster and nearly real-time system. This system can be incorporated into a larger surveillance system equipped with advanced algorithms involving anomalous event detection and activity recognition. This is a step towards more secure and robust surveillance systems and efficient video data analysis.

LogicaCMG were provided with an opportunity to deploy a facial recognition system in a realistic scenario. 12 cameras were installed at an international airport covering all entrances to the immigration hall. The evaluation took place over several months with numerous adjustments to both the hardware (i.e. cameras, servers and capture cards) and software. The learning curve has been very steep but a stage has now been reached where both LogicaCMG and the client are confident that, subject to the right environmental conditions (lighting and camera location) an effective system can be defined with a high probability of successful detection of the target individual, with minimal false alarms. To the best of our knowledge, results with a >90% detection rate, of non-compliant subjects 'at range' has not been achieved anywhere else. This puts this location at the forefront of capability in this area. The results achieved demonstrate that, given optimised conditions, it is possible to achieve a long range biometric identification of a non compliant subject, with a high rate of success.

Biometrics is the most emerging technology for automatic people authentication, nevertheless severe concerns raised about security of such systems and users' privacy. In case of malicious attacks toward one or more components of the authentication system, stolen biometric features cannot be replaced. This paper focuses on securing the enrollment database and the communication channel between such database and the matcher. In particular, a method is developed to protect the stored biometric templates, adapting the fuzzy commitment scheme to iris biometrics by exploiting error correction codes tailored on template discriminability. The aforementioned method allows template renewability applied to iris based authentication and guarantees high security performing the match in the encrypted domain.

In this paper we propose a signature-based biometric system, where watermarking is applied to signature images in order to hide and keep secret some signature features in a static representation of the signature itself. Being a behavioral biometric, signatures are intrinsically different from other commonly used biometric data, possessing dynamic properties which can not be extracted from a single signature image. The marked images can be used for user authentication, letting their static characteristics being analyzed by automatic algorithms or security attendants. When a higher security is needed, the embedded features can be extracted and used, thus realizing a multi-level decision procedure. The proposed watermarking techniques are tailored to images with sharpened edges, just like a signature picture. In order to obtain a robust method, able to hide relevant data while keeping intact the original structure of the host, the mark is embedded as close as possible to the lines that constitute the signature, using the properties of the Radon transform. An extensive set of experimental results, obtained varying the system's parameters and concerning both the mark extraction and the verification performances, show the effectiveness of our approach.

Information retrieval is critical in security technologies such as those for status identification and documentation authentication. Ideally, coding materials should be difficult to locate, impossible to counterfeit, and easy to process. This presentation addresses a novel information retrieval technology with these ideal features of its coding materials: the photo-luminescent (PL) quantum-dots (QD) synthesized via wet-chemistry approaches. As compared to traditional PL materials, they exhibit emission with narrower full width at half maximum, greater brightness, and higher photo-stability; also, their PL wavelength can be easily and accurately tuned via their size, structure, and composition. Due to such a feasible tune-ability, mainly, QDs have demonstrated enormous potential applications in security and defense. When QDs are excited, they can provide coded information with their PL wavelength and intensity. If the coding wavelengths from the QD PL are designed as the Fraunhofer lines, i.e. black lines in solar spectrum, the retrieval system can extract the useful information even under sunshine covering areas. Multi-photon excitation (MPE) technologies can further extend applications of QDs to multi-layer information extraction. For an info-label of 2-millimeter in depth, a MPE system with the depth resolution less than one micro-meter can thus achieve 2 GB resolutions, when a coding material exhibiting 6 PL wavelengths with 10 intensity levels. In general, transparent thin-film coating of QDs can be applied to various substrates, such as documents, fingernails, and military helmets and vehicles. Moreover, QD based security information can be easily destroyed by preset expiration in the presence of timing agents.

In this work we present a new way to mask the data in a one-user communication system when direct sequence - code division multiple access (DS-CDMA) techniques are used. The code is generated by a digital chaotic generator, originally proposed by us and previously reported for a chaos cryptographic system. It is demonstrated that if the user's data signal is encoded with a bipolar phase-shift keying (BPSK) technique, usual in DS-CDMA, it can be easily recovered from a time-frequency domain representation. To avoid this situation, a new system is presented in which a previous dispersive stage is applied to the data signal. A time-frequency domain analysis is performed, and the devices required at the transmitter and receiver end, both user-independent, are presented for the optical domain.

We propose a simple and robust method for the recovery of phase data pages. We provide experimental proof of the concept and investigate its applicability to optical encryption and encrypted holographic storage. Finally we discuss a possible compact optical implementation of the method.

The polarization-holographic system of protection has been developed. The suggested approach is based on new physical principles, namely on the using of polarization state of light that allows the level of protection of important documents, securities, industrial goods, etc. against counterfeiting to be increased. The system includes polarization-holographic protection elements and a device for the definition of authenticity of these elements. The elements are made on specially synthesized polarization-sensitive materials with certain properties which additionally raise the level of protection. The definition of authenticity is made by means of the analysis of the polarization state of light diffracted on the protective element. The essential advantage of this system is that the copying of the polarization-holographic protective element by optical methods is impossible in principle which complicates their counterfeiting.

In recent years, quantum cascade lasers (QCL) have been proven in robust, high-performance gas analyzers designed for continuous emission monitoring (CEM) in harsh environments. In 2006, Cascade Technologies reported progress towards adapting its patented technology for homeland security applications by publishing initial results on explosive compound detection. This paper presents the performance and results from a QCL-based people screening portal developed during the past year and aimed at the detection of precursors used in the make up of improvised explosive devices (IED). System tests have been carried out on a large number of potential interferents, together with target precursor materials, reinforcing original assumptions that compound fingerprinting can be effectively demonstrated using this technique. Results have shown that an extremely high degree of specificity can be achieved with a sub-second response time. Furthermore, it has been shown that unambiguous precursor signature recognition can be extended to compound mixtures associated with the intermediate stages in the make up of IEDs, whilst maintaining interferent immunity. The portal sensitivity was configured for parts per billion (ppb) detection level thresholds, but is currently being reconfigured for sub-ppb detection. In summary, the results obtained from the QCL based portal indicate that development of a low cost detection system, with enhanced features such as low false positive and high throughput screening of individuals or items, is possible. Development and testing was carried out with the support of the UK government.

Raman spectroscopy provides a very effective method of identifying an illicit substance in situ without separation or contact other than with a laser beam. The equipment required is steadily improving and is now reliable and simple to operate. Costs are also coming down and hand held portable spectrometers are proving very effective. The main limitations on the use of the technique are that it is insensitive in terms of the number of incident photons converted into Raman scattered photons and fluorescence produced in the sample by the incident radiation interferes. Newer methods, still largely in the development phase, will increase the potential for selected applications. The use of picosecond pulsed lasers can discriminate between fluorescence and Raman scattering and this has been used in the laboratory to examine street samples of illicit drugs. Surface-enhanced Raman scattering, in which the analyte requires to be adsorbed onto a roughened metal surface, creates a sensitivity to compete with fluorescence and quenches fluorescence for molecules on a surface. This provides the ability to detect trace amounts of substances in some cases. The improving optics, detection capability and the reliability of the new methods indicate that the potential for the use of Raman spectroscopy for security purposes will increase with time.

We report diffuse reflection imaging of concealed powdered samples in atmospheric air using a quantum cascade laser operating at 2.83 THz. The imaging system uses a helium-cooled silicon bolometer for mapping radiation diffusely reflected and scattered from samples, and a room-temperature pyroelectric sensor for simultaneously acquiring a specular image. A range of powders concealed within plastic packaging and standard FedEx envelopes was imaged with a resolution of better than 0.5 mm, and it was possible to detect powdered samples concealed within packaging from which there was a strong component of surface reflection. The feasibility of performing dual-wavelength diffuse reflection imaging for identification of illicit drugs and explosives is discussed.

There are situations where it is advantageous to visually obscure through glass, to an external observer, the movement of people within a well lit room. It may be that the building use has changed or existing measures which had provided obscuration such as 'Bomb-blast' curtains have been discontinued. Recognising that implemented solutions must create the minimum disruption to outward visibility and involve the least procedural effort (be simple to use), the Centre for Protection of National Infrastructure, CPNI, commissioned this study, defining key requirements including: (a) Automatic or simple manual operation (b) Obscuration of movement within the building from outside (c) Varying levels of obscuration depending on the difference in internal and external light levels. (d) Minimum disruption to outward visibility (e) Acceptable for use on heritage and iconic sites (f) Easy to retrofit (g) Low cost This report reviews earlier work carried out into the protection of Guardrooms by the use of lighting techniques coupled with the use of reflective and screen printed films. Other innovative solutions including Electrochromatic controllable glazing which may prove more appropriate to office and commercial buildings are also considered. It is seen that some measures, (window films or blinds), are cost effective and unsophisticated while more complex automatic systems using reactive glazing can offer critical design advantages. It must be noted however that some of the key requirements are mutually exclusive and any solution chosen will always be a compromise based on client needs and circumstances.

Lens technology has been developed that enables a scene to be scanned optically whilst the lens remains stationary. The covert optically-scanning enhanced (COSE) pinhole lens technology also has the option of incorporating a zoom capability. This allows any target in the scene to be isolated and zoomed onto for further scrutiny. The COSE lens has an external stationary entrance pupil. The operation of the lens may be visualized as optically pivoting about this pupil plane whilst scanning the scene but physically remaining stationary. A technology demonstrator has been built and its performance is discussed. Further developments of the COSE technology are also discussed.

This paper will describe the development of a real time matched filtering technique for hyperspectral imagery. Advanced multispectral, or hyperspectral, camera systems can potentially be used to identify objects of interest on the basis of spectral characteristics. Matched filtering relies on there being a measurable difference between the spectrum of the target and that of background materials such as soil, vegetation, concrete and tarmac. If prior knowledge is available then the target can be found by numerically matching the two spectra. Tests have been carried out to evaluate the effectiveness of a technique that utilises a specific type of spectral matched filter. Some results from the tests will be presented that indicate how our technique is affected by changes in environmental and illumination conditions. Example detections of a number of unclassified objects will be presented.

The report describes the results of a multi-year programme of research aimed at the development of an integrated multi-sensor perimeter detection system capable of being deployed at an operational site. The research was driven by end user requirements in protective security, particularly in threat detection and assessment, where effective capability was either not available or prohibitively expensive. Novel video analytics have been designed to provide robust detection of pedestrians in clutter while new radar detection and tracking algorithms provide wide area day/night surveillance. A modular integrated architecture based on commercially available components has been developed. A graphical user interface allows intuitive interaction and visualisation with the sensors. The fusion of video, radar and other sensor data provides the basis of a threat detection capability for real life conditions. The system was designed to be modular and extendable in order to accommodate future and legacy surveillance sensors. The current sensor mix includes stereoscopic video cameras, mmWave ground movement radar, CCTV and a commercially available perimeter detection cable. The paper outlines the development of the system and describes the lessons learnt after deployment in a pilot trial.

The range and scope of electro-optical (EO) sensor systems within security and surveillance (S&S) is growing, and this places a corresponding demand on the image processing functionality required to meet end users' requirements. Increasingly, these requirements include the ability to monitor wide areas with multiple affordable cameras, and for good quality imagery to be available 24-hours a day. This paper presents the results from some real-time systems which offer this capability, and focuses on a number of the image processing techniques used to deliver a high-quality, wide-angle, day/night capability. These include the production of a seamless image mosaic from multiple sensors, the removal of artefacts from scenes, enhancements to take account of changing environmental conditions, and a means of allowing the system to automatically focus on an area of interest highlighted by an operator. In addition, the cost of some high-performance S&S systems may be reduced by omitting physical calibration elements and performing sensor calibration using scene-derived data instead, and so a method for achieving this is also reported. The paper considers both the theoretical aspects of the algorithms presented and the issues involved in real-time implementation for S&S applications.

In this paper we propose an OFDM (Orthogonal Frequency Division Multiplexing) wireless communication system that introduces mutual authentication and encryption at the physical layer, without impairing spectral efficiency, exploiting some freedom degrees of the base-band signal, and using encrypted-hash algorithms. FEC (Forward Error Correction) is instead performed through variable-rate Turbo Codes. To avoid false rejections, i.e. rejections of enrolled (authorized) users, we designed and tested a robust hash algorithm. This robustness is obtained both by a segmentation of the hash domain (based on BCH codes) and by the FEC capabilities of Turbo Codes.