Face re-identification is a challenging task which is aimed to check similarity of two faces shown in the images. Face recognition system have been investigated since many years mostly in visible domain. We investigate face recognition methods based on facial images acquired in far-infrared range (thermal spectrum). The main reason for using thermal infrared for face recognition is to observe people in night conditions. However, this task is not free of challenges. In this paper we investigate the impact of various head positions on efficiency of face re-identification. The paper presents our measurement approach, results of many series of tests as well as performance metrics of re-identification based on three state-of-the-art facial descriptors.
The paper describes the results of experimental research on the mobile verification of travellers based on fingerprints. Three-day tests were carried out at the border crossing in Terespol, Poland. The developed system automatically acquires personal and biometric data (fingerprints) from the Polish biometric passport, determines their quality and compares with the live data collected from the traveller. In addition, the system measures the time of individual stages of the process and determines total transaction time. For total number of correctly scanned travellers equal to 128, false acceptance rate equals to 0, while rejection rate is less than 1%. The average transition time of border check was 37 seconds.
Biometrics refers to unique human characteristics. Each unique characteristic may be used to label and describe individuals and for automatic recognition of a person based on physiological or behavioural properties. One of the most natural and the most popular biometric trait is a face. The most common research methods on face recognition are based on visible light. State-of-the-art face recognition systems operating in the visible light spectrum achieve very high level of recognition accuracy under controlled environmental conditions. Thermal infrared imagery seems to be a promising alternative or complement to visible range imaging due to its relatively high resistance to illumination changes. A thermal infrared image of the human face presents its unique heat-signature and can be used for recognition. The characteristics of thermal images maintain advantages over visible light images, and can be used to improve algorithms of human face recognition in several aspects. Mid-wavelength or far-wavelength infrared also referred to as thermal infrared seems to be promising alternatives. We present the study on 1:1 recognition in thermal infrared domain. The two approaches we are considering are stand-off face verification of non-moving person as well as stop-less face verification on-the-move. The paper presents methodology of our studies and challenges for face recognition systems in the thermal infrared domain.
Biometrics is a technique for automatic recognition of a person based on physiological or behavior characteristics. Since the characteristics used are unique, biometrics can create a direct link between a person and identity, based on variety of characteristics. The human face is one of the most important biometric modalities for automatic authentication. The most popular method of face recognition which relies on processing of visual information seems to be imperfect. Thermal infrared imagery may be a promising alternative or complement to visible range imaging due to its several reasons. This paper presents an approach of combining both methods.
Biometrics is a science that studies and analyzes physical structure of a human body and behaviour of people. Biometrics found many applications ranging from border control systems, forensics systems for criminal investigations to systems for access control. Unique identifiers, also referred to as modalities are used to distinguish individuals. One of the most common and natural human identifiers is a face. As a result of decades of investigations, face recognition achieved high level of maturity, however recognition in visible spectrum is still challenging due to illumination aspects or new ways of spoofing. One of the alternatives is recognition of face in different parts of light spectrum, e.g. in infrared spectrum. Thermal infrared offer new possibilities for human recognition due to its specific properties as well as mature equipment. In this paper we present the scheme of subject’s verification methodology by using facial images in thermal range. The study is focused on the local feature extraction methods and on the similarity metrics. We present comparison of two local texture-based descriptors for thermal 1-to-1 face recognition.
Multispectral systems for detection of concealed dangerous objects are becoming more popular because of their higher effectiveness compared to mono-spectral systems. So far, the problem of detecting objects hidden under clothing was considered only in the case of airports but it is becoming more important for public places like metro stations, and government buildings.
Exploration of new spectral bands as well as development of technology result in introduction of new solutions – both mono and multispectral. It has been proved that objects hidden under clothing can be detected and visualized using terahertz (THz) cameras. However, passive THz cameras still offer too low image resolution for objects recognition. Limited range is another issue of passive imagers. On the other hand new infrared cameras offer sufficient parameters to detect objects covered with fabrics in some conditions, as well as high image quality and big pixel resolutions.
The purpose of the studies is to investigate and compare the possibilities of using passive cameras operating in long wavelength infrared (LWIR) and THz spectral ranges for detection of concealed objects. For the purpose of investigations, commercial imagers operating in 6.5-11.7 μm and 250GHz (1.25mm) were used. In the article, we present the measurement setup and the results of measurements in various operating conditions. Theoretical studies of both spectral bands focused on detection of objects with passive imagers are also presented.
Terahertz radiation is within the frequency range from 100 GHz to 10THz. This radiation has specific characteristics in
terms of imaging. The radiation is harmless to the human body because the energy transferred by electromagnetic waves
in this range of frequencies are very small thus there is no ionization of matter.
The development of imaging devices and exploration of new spectral bands is a chance to introduce new equipment for
assuring public safety. It has been proved that objects hidden under clothing can be detected and visualized using
terahertz (THz) cameras. However, passive THz cameras still offer too low image resolution for objects recognition.
In order to determine the properties of terahertz imaging for detection of hidden objects several aspects need to be
considered. Taking into account the fact that the image captured by the terahertz camera reflects the spatial distribution
of the relative temperature of the observed objects, the effect of the measurement time on the imaging capabilities should
be examined. A very important aspect is the influence of the type (material composition) of coating material, as well as
the type of an object hidden under clothing (size and material).
The purpose of the studies is to investigate the time stability of passive THz imaging on 250 GHz for detection of
concealed objects. In the article, we present the measurement setup, the measurement methodology as well as the initial
results of measurements with various types of clothing and test objects.
Risks to the safety of public zones (generally available for people) are related mainly to the presence of hidden dangerous objects (such as knives, guns, bombs etc.) and their usage. Modern system for the monitoring of such zones attempt to detect dangerous tools using multispectral cameras working in different spectral ranges: the visible radiation, near, medium and long range infrared and recently also in terahertz range. In order to develop methods and algorithms to detect hidden objects it is necessary to determine the thermal signatures of such objects of interest. The laboratory measurements were conducted to determine the thermal signatures of dangerous tools hidden under various clothes in different ambient conditions. Cameras used for measurements were working in spectral range 0.6-12.5 µm. An infrared imaging Fourier transform spectroradiometer was also used, working in spectral range 7.7-11.7 µm. Analysis of registered thermograms and hyperspectral datacubes has yielded the thermal signatures for: two types of guns, two types of knives and home-made explosive bombs. The determined thermal signatures will be used in the development of method and algorithms of image analysis implemented in proposed monitoring systems.
Detection of concealed dangerous objects is a very demanding problem of public safety. So far, the problem of detecting
objects hidden under clothing was considered only in the case of airports but it is becoming more and more important for
public places like metro stations, and government buildings.
The development of imaging devices and exploration of new spectral bands is a chance to introduce new equipment for
assuring public safety. It has been proved that objects hidden under clothing can be detected and visualized using
terahertz (THz) cameras. However, passive THz cameras still offer too low image resolution for objects recognition. On
the other hand new infrared cameras offer sufficient parameters to detect objects covered with fabrics in some
conditions, as well as high image quality and big pixel resolutions.
The purpose of the studies is to investigate the possibilities of using various cameras operating in different spectral
ranges for detection of concealed objects. In the article, we present the measurement setup consisting of medium
wavelength infrared (MWIR), long wavelength infrared (LWIR), THz and visible cameras and the initial results of
measurements with various types of clothing and test objects.
A growing interest in terahertz technology finds support in a large number of applications. One of the most interesting applications of terahertz waves is imaging. The terahertz range of electromagnetic radiation has large potential in the field of hidden objects detection because it is not harmful to humans [1, 2]. However, the main difficulty in the THz imaging systems is low image quality due to low sensitivity and a small number of pixels in detecting modules of cameras Considering the fact that even THz images with low pixel resolution still provide valuable information, it is justified to combine them with the high-resolution images from a visible camera. Image fusion can be used in a wide range of security applications for example detection and identification of hidden objects. Our goal is to build a system harmless to humans for screening and detection of hidden objects using a THz camera. A very important aspect of applying various processing techniques to images is proper assessment of image quality. We propose a combination of two image quality assessment methods (IQA) as a methodology of assessing quality of the fused images and a method to compare image fusion algorithms.
Terahertz imaging, is the latest entry into the crowded field of imaging technologies. Many applications are emerging for the relatively new technology. THz radiation penetrates deep into nonpolar and nonmetallic materials such as paper, plastic, clothes, wood, and ceramics that are usually opaque at optical wavelengths. The T-rays have large potential in the field of hidden objects detection because it is not harmful to humans. The main difficulty in the THz imaging systems is low image quality thus it is justified to combine THz images with the high-resolution images from a visible camera. An imaging system is usually composed of various subsystems. Many of the imaging systems use imaging devices working in various spectral ranges. Our goal is to build a system harmless to humans for screening and detection of hidden objects using a THz and VIS cameras.
Visible, LLTV and anti-fog cameras have different abilities to capture external information, and combining these abilities of the three cameras can greatly improve the environment perception ability of intruders or emergency situations. Designed specifically for professional surveillance use, multi active pixel sensor camera of this type allows targets to be monitored at long distance and tracked using the proportional pan, tilt and zoom system. In the article the research results of the camera to determine realistic and standardized parameters, e.g. range of detection, recognition and identification of humans are described. The paper presents measuring equipment, procedures and results.
Compressed sensing as an imaging method has become very popular among scientists and is becoming more and more
popular among hardware manufacturers. There are many hardware variants of single-pixel compressed sensing based
camera and there are many algorithms of sparse signal approximations. This fact makes it appear more and more
applications of compressive imaging.
Recently, many algorithms for signal reconstruction have been developed, however, all of them need many parameters to
be properly set before using. Setting proper parameters is crucial for preparing a real model of the single pixel camera as
well as for fast and efficient image synthesis. Because of high complexity of image recovery algorithms, image synthesis
process needs to be optimized.
Optimization of signal acquisition and processing parameters can be achieved running various camera simulations. In
the paper we present the integrated test environment for image synthesis of the single pixel camera and the test results of
simulations run with various configurations and parameters values. We used two combined adaptive methods for image
reconstruction - the Newton method and the conjugate gradient method. Test environment allows to run two kinds of
tests. The first test type is simulation of various parameters of acquired signal e.g. bit resolution. Image geometric
transformation like rotation is the second type of tests. Simulation results include quality parameters values of MSE,
PSNR and SSIM and image reconstruction time. Integrated test environment can be used during the process of hardware
selection as well as during camera tests with real signals.
We demonstrate the improvement of the quality of the image captured by TS4 - the commercially available THz passive
camera manufactured by ThruVision Systems Ltd. The measurements range of this device reaches 10 meters. Our
approach is based on application of novel spatial filters and algorithms, developed by us for computer processing of
passive THz images produced by the various THz cameras.
In our opinion, the most important result of this paper consists in a demonstration of the possibility of using a passive
THz camera to observe a difference in temperature on the human skin if this difference is caused by different
temperatures in the inside of the body. Such possibility was proposed by Vyacheslav Trofimov on the Conference in
Baltimore (April 2012) as well as and in . We discuss two physical experiments, in which a person drinks hot and
cold water. After computer processing of images captured by passive THz camera TS4 we may see the pronounced
temperature trace on the human body. We illustrate this phenomenon by a series of images captured by passive THz
camera in real time.
As we believe, these experiments allow us wide applications of passive THz cameras for the detection of objects
concealed in the inside of the human body because the difference in temperature that will be reflected on the human
skin. Modern passive THz cameras have not enough resolution in temperature to see this difference. However, computer
processing allows us to enhance it for this application.
Using computer processing one may enhance the image quality and delete noise on the images. In some cases, it is
possible to achieve full de-noising of the image.
Terahertz technology is one of emerging technologies that has a potential to change our life. There are a lot of attractive applications in fields like security, astronomy, biology and medicine. Until recent years, terahertz (THz) waves were an undiscovered, or most importantly, an unexploited area of electromagnetic spectrum. The reasons of this fact were difficulties in generation and detection of THz waves. Recent advances in hardware technology have started to open up the field to new applications such as THz imaging. The THz waves can penetrate through various materials. However, automated processing of THz images can be challenging. The THz frequency band is specially suited for clothes penetration because this radiation does not point any harmful ionizing effects thus it is safe for human beings. Strong technology development in this band have sparked with few interesting devices. Even if the development of THz cameras is an emerging topic, commercially available passive cameras still offer images of poor quality mainly because of its low resolution and low detectors sensitivity. Therefore, THz image processing is very challenging and urgent topic. Digital THz image processing is a really promising and cost-effective way for demanding security and defense applications. In the article we demonstrate the results of image quality enhancement and image fusion of images captured by a commercially available passive THz camera by means of various combined methods. Our research is focused on dangerous objects detection - guns, knives and bombs hidden under some popular types of clothing.
Solution presented in this article is a system using image acquisition time gating method. The time-spatial framing
method developed by authors was used to build Laser Photography System (LPS). An active vision system for open
space monitoring and terrorist threats detection is being built as an effect of recent work lead in the Institute of
Optoelectronics, MUT. The device is destined to prevent and recognize possible terrorist threats in important land and
marine areas. The aim of this article is to discuss the properties and hardware configuration of the Laser Photography
Screening cameras working in millimetre band gain more and more interest among security society mainly due to their
capability of finding items hidden under clothes. Performance of commercially available passive cameras is still limited
due to not sufficient resolution and contrast in comparison to other wavelengths (visible or infrared range). Testing of
such cameras usually requires some persons carrying guns, bombs or knives. Such persons can have different clothes or
body temperature, what makes the measurements even more ambiguous. To avoid such situations we built a moving
phantom of human body. The phantom consists of a polystyrene manikin which is covered with a number of small pipes
with water. Pipes were next coated with a silicone "skin". The veins (pipes) are filled with water heated up to 37 C
degrees to obtain the same temperature as human body. The phantom is made of non-metallic materials and is placed on
a moving wirelessly-controlled platform with four wheels. The phantom can be dressed with a set of ordinary clothes and
can be equipped with some dangerous (guns, bombs) and non-dangerous items. For tests we used a passive commercially
available camera TS4 from ThruVision Systems Ltd. operating at 250 GHz. We compared the images taken from
phantom and a man and we obtained good similarity both for naked as well as dressed man/phantom case. We also tested
the phantom with different sets of clothes and hidden items and we got good conformity with persons.
The recent development of electro-optical instrumentation allowed constructing 4D (3D + time) structure-light scanners
which may be used to measure the surface of human body in motion. The main advantage of structure-light scanners is
the possibility of capturing data from the whole measured body surface, while traditional marker-based systems acquire
data only form markers attached to skin of the examined patient. The paper describes new parameters describing the local shape of measured surface. The distribution maps of these
parameters allow discrimination of various surface types and in effect localization and tracing of under-skin anatomical
structures in time. The presented parameters give similar results to well-known curvatures but are easier and quicker to
calculate. Moreover the calculation process of the new parameters is more numerically stable itself. The developed path of processing and analysis of 4D measurement data has been presented. It contains the following
stages: data acquisition, volumetric model creation, calculations of shape parameters, selecting areas of interest, locating
and tracing of anatomical landmarks. Exemplary results of application of developed parameters and methods to real measurement and computer generated
data are also presented.
Orthopedics and neurosciences are fields of medicine where the analysis of objective movement parameters is extremely important for clinical diagnosis. Moreover, as there are significant differences between static and dynamic parameters, there is a strong need of analyzing the anatomical structures under functional conditions. In clinical gait analysis the benefits of kinematical methods are undoubted.
In this paper we present a 4D (3D + time) measurement system capable of automatic location of selected anatomical structures by locating and tracing the structures' position and orientation in time. The presented system is designed to help a general practitioner in diagnosing selected lower limbs' dysfunctions (e.g. knee injuries) and also determine if a patient should be directed for further examination (e.g. x-ray or MRI).
The measurement system components are hardware and software. For the hardware part we adapt the laser triangulation method. In this way we can evaluate functional and dynamic movements in a contact-free, non-invasive way, without the use of potentially harmful radiation. Furthermore, opposite to marker-based video-tracking systems, no preparation time is required.
The software part consists of a data acquisition module, an image processing and point clouds (point cloud, set of points described by coordinates (x, y, z)) calculation module, a preliminary processing module, a feature-searching module and an external biomechanical module.
The paper briefly presents the modules mentioned above with the focus on the feature-searching module. Also we present some measurement and analysis results. These include: parameters maps, landmarks trajectories in time sequence and animation of a simplified model of lower limbs.