rus
Issue #8/2023
A. S. Yusupovsky, V. V. Grishachev
Detection of Covert CCTV Systems Based on the Smartphone Depth Sensor
Detection of Covert CCTV Systems Based on the Smartphone Depth Sensor
DOI: 10.22184/1993-7296.FRos.2023.17.8.638.655
The paper presents research materials related to the use of a smartphone video system (depth sensor) to detect the covert video cameras in various premises, including the places of temporary accommodation (hotel rooms, locker rooms, passenger compartments, etc.) where a person suspects an intrusion on his personal security or privacy breach. The depth sensor of a smartphone makes it possible to increase the efficiency of automated search for covert video cameras by optical flare for an untrained person and security specialists. The article formulates proposals for the search procedure for covert video cameras and identifies the main technical specifications of an efficient survey.
The paper presents research materials related to the use of a smartphone video system (depth sensor) to detect the covert video cameras in various premises, including the places of temporary accommodation (hotel rooms, locker rooms, passenger compartments, etc.) where a person suspects an intrusion on his personal security or privacy breach. The depth sensor of a smartphone makes it possible to increase the efficiency of automated search for covert video cameras by optical flare for an untrained person and security specialists. The article formulates proposals for the search procedure for covert video cameras and identifies the main technical specifications of an efficient survey.
Теги: covert video camera personal information security smartphone depth sensor visual information визуальная информация датчик глубины смартфона информационная безопасность личности скрытая видеокамера
Detection of Covert CCTV Systems Based on the Smartphone Depth Sensor
A. S. Yusupovsky, V. V. Grishachev
Russian State University for the Humanities, Moscow, Russia
The paper presents research materials related to the use of a smartphone video system (depth sensor) to detect the covert video cameras in various premises, including the places of temporary accommodation (hotel rooms, locker rooms, passenger compartments, etc.) where a person suspects an intrusion on his personal security or privacy breach. The depth sensor of a smartphone makes it possible to increase the efficiency of automated search for covert video cameras by optical flare for an untrained person and security specialists. The article formulates proposals for the search procedure for covert video cameras and identifies the main technical specifications of an efficient survey.
Keywords: personal information security, visual information, covert video camera, smartphone depth sensor
Article received: 17.09. 2023
Article accepted: 07.11. 2023
INTRODUCTION
Protection of information related to the personal data in the form of photo/video materials is an important element in ensuring the private space security. Any personal information security threats can occur in various situations. One of such cases is location of a person in his private space, when he must be sure of protection against any covert surveillance, for example, being in a hotel room, in public places with limited visual access, etc. Such threats can be of great importance for the personal information security [1, 2] and result, for example, in the form of unauthorized publishing of obtained photo/video materials on social networks and other web resources.
The issue of information security of a personal private space shall be resolved not only by the state in the form of passage of various laws, provision of special services, etc., but also by the technical capabilities of the citizen. The technical devices that ensure personal information security must be efficient, economically and technically accessible to every person. In the contemporary society, such a device is a smartphone performing a wide range of functions from any communication to entertainment capabilities. The smartphones have a set of sensors that can be used, among other things, to ensure personal information security.
While the security services with powerful technical and legal capabilities are responsible for ensuring the visual information protection at the state level and at the level of large-scale companies, then confidentiality for an individual or a small company shall be maintained exclusively by personal precautions and allowable generally available technical devices. Such restrictions lead to the search for the most efficient and affordable technical solutions that do not burden a person financially or in other ways.
Technical intelligence and visual information protection
The covert video surveillance in the field of technical intelligence can be performed using video cameras for various purposes [2, 3]: from any custom web cameras to the special covert video cameras camouflaged as the ordinary household items. The small-sized pinhole cameras with an extended pupil, a typical lens size of less than a millimeter, and conversion of an optical image into an electronic or digital signal in the silicon CCD matrices and open (broadcasting) or closed (cable) transmission to the recording system are used in the indoor covert CCTV systems at the short distances (up to tens of meters).
The countermeasures against the visual intelligent systems using the covert CCTVs can be divided into two types, related to their functioning:
Covert video surveillance countering measures based on the detection of electromagnetic radiation (EMR) of a running video camera [2, 5]; by searching for the electronic elements by nonlinear location using electromagnetic radiation; as well as neutralizing the system’s electronics by electromagnetic interference or destruction by an electromagnetic pulse. This method has its own limitations related to performance of the activities using special equipment that is not always available.
The Protect K18 detector (Fig. 1) is commercially available. It allows scanning of the air EMR within the range from 1 MHz to 8 GHz and identifying radio communication channels with the external devices with a sensitivity of more than 30 μW and a declared range of 0.1 m to 15 m. The device is small-sized, but does not help to identify non-emitting devices with the internal media recording or transmissions via the cable systems. The application of such systems in the urban environments is limited by the difficult radio environment in the surrounding area.
Covert video surveillance countering measures based on destructive effects on the optical system through significant flash exposure in the optical or infrared region of the EMR spectrum [2]. This method requires either preliminary detection of the surveillance system optical devices, or isotropic flash exposure of the entire space that needs the comprehensive provision of special equipment. The most efficient method for identifying covert video surveillance systems is based on the flares developed when optical radiation is reflected from the objective lenses and CCD matrix.
Optical system of a covert video camera and its detection
The basis of any CCTV system is an optical system forming an image of the surrounding space on a CCD matrix [2, 3].
The main optical system for covert video surveillance are the cameras with a pinhole lens (Fig. 2), i. e. a lens with a pupil (1) remote at a distance l with a diameter d (less than 1 mm) and typical dimensions of several mm, determined by the focal length f of the lens (2). The input optical radiation is focused on a CCD matrix (3) that converts the optical signal into an electrical signal (4). The incoming light flux at the viewing angle ϕ is subject to the Fresnel reflections on the focusing lenses of the objective and matrix (5) that are perceived as the optical flares, i. e. bright light spots on a uniform background in the visible and other regions of the optical spectrum. Such flare detection in the visible spectrum is quite difficult.
As an example, it is possible to use a laptop camera under the protective glass (Fig. 3). In this case, the video camera can be detected in the visible light by the design features of protective shutter and reflective coating around the camera pin-hole (dark dot in the light outline on the right figure). In the case of oblique shooting, the pin-hole is not visible, but in the case of ordinary light incidence it is difficult to distinguish against a dark background for the human eye, even it is located at the close distance (a laptop). To reduce the Fresnel losses related to the reflections, the camera’s optical elements are coated by the antireflection films that reduce reflection by 3–4 times in comparison to the standard value of 4% [6]. In the case of ordinary light, such surfaces form weak reflections and are therefore difficult to detect with the eye. The antireflection coating relates to the central region of the visible spectrum and, to a lesser extent, visible red light and, even to a lesser extent, invisible infrared light. It is possible to increase observability by moving from the visible region to the visible red spectrum part, by illuminating with intense red light and observing the area with a video camera through a red color filter.
The simplest detection system for video camera lenses includes the devices such as the Protect K18 detector (Fig. 1) that has several LED red light sources. When observing a surface irradiated with the red light through a red color filter in the detector’s center, the optical surfaces form flares visible to the eye (if available). This method depends on the experience of an examiner and requires meticulous long-term work when examining the large surfaces. Any professional cover video camera detectors apply more powerful monochrome LED light sources with a small viewing angle of several degrees and an electronic light detection system that make it possible to register reflective objects at the distances of tens of meters.
Transition to the invisible infrared portion of the spectrum increases the Fresnel reflection that is not blocked by antireflection coatings. The brightness of flares, invisible to the eye, is increased, therefore, its observation requires application of the infrared radiation receivers.
Smartphone video system
and depth sensor
The up-to-date wearable devices, such as smartphones, can provide an affordable and efficient solution to protect individuals against any visual reconnaissance [7]. A smartphone is a generally accessible device with a wide range of functionality. Firstly, it is a communication device that allows perform local communication using Bluetooth technology, Wi-Fi and long-distance communication using 4G/5G technology, etc. Secondly, the smartphones include a large number of sensors from the sensors converting audio and video information to the physical field sensors. All sensors can be built-in or external with a wired or wireless connection. Any smartphone can be used as a computing device to measure any physical quantities using its own or connected sensors, the capabilities of which are limited only by the power supplies and computational resources [7].
One of the advanced smartphone systems is its video system for visual information recording and display. For example, Samsung Galaxy A22 (Fig. 4) has five video cameras for various purposes: a front-facing camera for selfies, a macro (close-up) camera, an ultra-wide-angle camera for shooting the close-up large objects or panoramic shooting, a main camera for the high-quality portrait photographs, a depth camera for sharpening the frame and other purposes. In addition to the cameras noted, there are some built-in cameras for night photography, black and white cameras and others that can be of external nature, such as the thermal imaging cameras.
In smartphones, 3D images are created using the stereoscopic vision systems based on two cameras (stereo vision); structural light projection by infrared radiation in the form of a patterned set of points and lines distributed in the space, the curvature of which determines the three-dimensional object structure (structured light). Recently, the iPad Pro 2020 mobile devices are provided with the lidar systems. The depth sensors are the most efficient option for creating a three-dimensional image (ToF sensors, time-of-flight cameras, ToF cameras, depth cameras) to be used to measure distance, recognize gestures and faces, develop 3D object models, proximity sensor and for other purposes [8, 9, 10].
The depth sensor (ToF-camera) is based on measuring the propagation time of a light pulse from an infrared laser to a reflective obstacle and back to a detector in the form of an IR camera (Fig. 5). The IR laser (2) of the depth sensor (1) generates the modulated optical radiation directed at the object under study being synchronized with the IR camera operation process (3). The optical radiation (4, 5) reflected from various objects (6, 7, 8) is recorded by an IR camera with a time delay. The duration of exposure to the laser radiation is several nanoseconds that makes it possible to record the phase shift of reflected radiation with the millimetric accuracy. When processing the delay time (time of flight), a 3D image of the probed space is generated by isolating the areas with the same delay time to be highlighted in brightness and color. Various shades of red light are usually used: the closer areas are brighter, and the more distant ones are dimmer.
The 3D image generation by the depth sensor camera depends on the IR laser intensity and the IR camera sensitivity; the higher the radiation intensity, the greater the reflected signal intensity and the less error of the distance to be determined. As the distance of the object under study is increased, intensity of the reflected IR radiation decreases leading to the enhanced measurement error. On the other part, as the distance to the object is decreased, the time of flight is reduced and the distance measurement error is also increased. Consequently, each depth sensor has its own region of space in terms of a distance at which it operates most efficiently.
The efficiency of a response generation from objects in the form of holes with a small reflective diameter and shallow depth, such as a covert camera with a pinhole lens, depends on the field of view (viewing angle) of the hole (lens). Only when the laser radiation enters the field of view of the lens, the response is formed that imposes another limitation for observing such objects, namely the viewing angle of the observed covert camera hole and the source of laser radiation. Displacement of the depth sensor along the surface under study must be within the viewing angles of the lens and laser.
Any proposals for using a depth sensor to detect a covert camera are given in some papers [10]. In particular, it is proposed to develop software for installation on a smartphone with a depth sensor to detect any images obtained from the small holes in the form of flares, i. e. significant small reflections, clearly visible against a generally uniform background. Such papers do not discuss connection with the technical parameters of a smartphone, so our paper focuses on capabilities of the smartphone’s depth sensor and possible practical inspection of premises for the availability of covert cameras.
Practical simulation of video camera detection using a smartphone’s depth sensor
The practical studies of detecting covert video cameras were performed using a Samsung Galaxy S 20+ smartphone with a built-in DeethVision ToF-camera (Fig. 4) [11]. The depth sensor operates in the ambient imaging mode using the open source ToF Viewer Android application [12]. It shows the resulting video image on the smartphone display in real time. The DeethVision camera functions in the invisible IR range with its own source that accounts for the maximum sensitivity of the silicon CCD matrix. The generated depth map is perfectly visible in absolute darkness.
The practical studies consisted of demonstration of the possible camera detection by the flare of optics, determination of technical specifications of the depth sensor of a given smartphone, for which a convert video camera model was prepared. On the basis of experimental studies performed, several proposals were made for practical implementation of the covert CCTV detection process.
Demonstration of the ability to detect video cameras using a smartphone’s depth sensor
Verification of the possible camera detection using a smartphone’s depth sensor was carried out by scanning of the external web camera (Fig. 6) and a laptop web camera (Fig. 7). The photographs given were taken with a conventional camera (on the left) and a camera with a depth sensor (on the right) at a normal location at various distances (farther and closer). As one can see from the photographs in normal light, even when placed closely, it is not possible to distinguish the webcam lens of both the external and built-in laptop cameras. It is possible to determine the camera existence by the light-colored rings around the lens entrance hole typical for the web cameras. These differences are more clearly visible during the eyeball inspection when the viewing angle is changed. When approaching the external webcam, it is possible to distinguish the pinhole lens. When using a camera with a depth sensor, the pinhole lens is clearly visible when viewed closely and is difficult to see when moving away from the object. Thus, it is possible to confirm a covert video camera detection process based on the flares from a pinhole lens using a camera with a depth sensor.
Test model of a covert video camera
A WI-FI camera was used as a control video camera simulating the covert video surveillance system (Fig. 8). In order to increase the external covertness attributes by reducing the physical dimensions, the optical unit (lens with a CCD matrix in the housing) was removed from the camera. This activity reduced the camera dimensions while completely disrupting its functionality. The physical parameters of the optical unit were determined by the hole size for the input optics that was about 3 mm, with a total diameter of the cylindrical part of the camera optical unit of 20 mm and a thickness of 24 mm.
The camera was camouflaged to match the external background using a black sheet of paper, where a hole was made for the lens (Fig. 9). The black paper background faded with the camera’s black color that simulated the camera placement secrecy. During the simulation process, the task of preserving the shooting functions and completely concealing the camera placement was not set, i. e. only the conditional simulation of functioning and placement of a covert video camera was provided.
All subsequent experimental studies were performed with the described simulated covert camera in the form of a lens inserted into a hole in a black sheet of paper (Fig. 8 and Fig. 9).
Experimental determination of technical capabilities of the depth sensor in a Samsung Galaxy S 20+ smartphone
The test objective was to determine the technical parameters for a simulated covert video camera detection using a camera with a smartphone depth sensor, such as the optimal distance to the object and the maximum viewing angle at which optical flare was observed most clearly.
The optical circuit of the experimental studies is shown in Fig. 10. The covert camera model (1) against the black screen (2) was placed at a certain distance l opposite to the smartphone camera (3) and the optical flare was recorded. The smartphone approached the object at a minimum distance and then moved away until the flare disappeared completely at a distance lmax. As a result of this test, the maximum observation distance lmax of a covert camera based on the flare was recorded with the smartphone in normal position. Further, the camera was placed at a distance lmax / 2 being half the maximum value for optimal observation of flares. At this distance, the smartphone moved parallel to the screen at a distance lplus, when the flares discontinued according to which the camera viewing angle ϕt was determined. All results were documented by photographing a model at experimentally determined distances with a conventional camera and a smartphone ToF-camera with visualization using the ToF Viewer application.
At the beginning of the experiment, a distance was selected equal to 40 cm (Fig. 11). With the normal smartphone position and a conventional camera, the covert camera optics were almost not observed. However, the ToF-camera provided its clear observation. The photographs made by the ToF-camera show not only the direct flares, but also the flares obtained when reflected from the table that complicates the task of identifying informative flares using the computer methods.
During the next experiment (Fig. 12), the studies were performed at a distance of 85 cm, when a conventional camera did not allow flares from the covert optics to be seen, and the ToF-camera generated a completely visible image. Moreover, as can be seen in the photographs, additional flares related to the reflections from auxiliary surfaces are not observed, since the side reflected light fluxes of the IR laser do not fall into the field of view of the covert camera and do not generate additional flares as in the previous experiment (Fig. 11). Based on this experiment, it is possible to estimate the limiting angle (about 15° × 2) of the viewing field of a model covert camera as the arctangent of the flare height above the table and the table diameter.
When moving at a distance of 120 cm, it is still possible to observe the flares, but when the smartphone is removed at a distance of more than 140 cm from the model, it leads to a complete loss of flare observation using the ToF-camera (Fig. 13). Thus, the test results confirm that the observation of flares is stopped at a distance lmax = 140 cm. The next stage of the research, namely determination of the camera’s field of view angle, was carried out at a distance of 70 cm from the object with parallel movement of the smartphone relative to the screen. As a result of measurements, the obtained estimates of the field of view limiting angle (Fig. 12) of 15x2 degrees were confirmed.
Experimental comparison of technical capabilities of the depth sensor in a Samsung Galaxy S 20+ smartphone and wiretapping detector Protect K‑18S
As it has been noted above, to search for covert optics, the K‑18 detector can be used. It has a visible red light source and a red optical filter to observe the reflected red light through the eyepiece.
A comparison of capabilities of the ToF-camera and the K18 sensor in detecting the lens flares demonstrates the advantages of the first method when inspecting the rooms for the covert camera availability using the lens flares. The results of experimental comparison of the sensors are given in Fig. 14 and fig. 15. As it can be seen from the photographs, the flares recorded by the K‑18 sensor are much weaker than provided by the ToF-camera. In the comparison experiment (Fig. 14 and Fig. 15), the image generated by the K‑18 sensor was recorded by the smartphone camera that approximated the object and had a higher sensitivity to red light than the human eye. In reality, observation must be performed only with the eye through the lens that further complicates the search procedure.
Study results and proposals for detecting covert video cameras
The practical studies performed and the results obtained make it possible to propose the technology and method for conducting inspections of premises and confidence rooms for the availability of covert CCTV systems in order to protect the personal private space. A smartphone with a depth sensor can be used to professionally protect visual information while improving its performance. The detection efficiency of the convert video cameras can be enhanced by increasing power and coherence (monochrome) of the infrared laser radiation that will lead to an increase in the detection range of cameras, due to the higher reflected power, and measurement accuracy of the difference in distances traveled by the laser radiation as a result of phase measurement methods. In practice, to detect any flares from the camera optics, the CCD matrix and smartphone display quality (resolution, brightness, contrast) is of great importance. The given recommendations can significantly increase the detection efficiency of professional covert video cameras with a small input hole (pinhole) of the lens that can be less than 1 mm while developing low intensities of reflected radiation.
Based on the research results, it is possible to identify the necessary inspection conditions for searching for video cameras:
Based on these requirements, several real time room inspection methods can be proposed.
A. Room inspection method (including confidence rooms) using a smartphone with a depth sensor to detect covert video cameras
Setup/preparation/calibration. While using standard available video cameras, configure a smartphone with a depth sensor for detection: determine the type of flares, establish the maximum observation length lmax and distance lplus;
Conduct an inspection of the room walls with image recording, placing the smartphone camera at the optimal distance lmax/2 and moving it with swinging parallel to the wall;
Visually identify the hotspots suspected for the availability of video cameras based on the flares occurred;
Display the video recording on a large monitor screen in the slow-motion mode: visually re-identify the hotspots suspected for the availability of video cameras based on the flares occurred;
Separately examine the hotspots at a closer distance with slow movement of the smartphone’s video camera;
Draw conclusions about the availability of covert video cameras.
When the person examining the premises does not have significant experience or does not have the physical capabilities for an effective survey, such as, for example, poor vision, poor attention to details, etc., he can involve any remote assistants in the examination or conduct the examination in the online mode using the smartphone’s communication capabilities. In this case, the method will be changed while maintaining all the basic principles. However, the process efficiency and velocity can be increased.
B. Room inspection method by a smartphone with a depth sensor to detect covert video cameras in the online mode with an assistant
Prepare a smartphone for detection of video cameras using a depth sensor by the example of proposals from the previous methods;
Establish an online connection with a remote assistant for real time transmission of video images to a remote large monitor;
Monitor the premises with the constant voice communication with an assistant.
At the best case, the assistant is a specialist or expert in the field of covert camera detection who is located at a distance and provides the real time support and advice. He helps during the examination for availability of covert video cameras, analyzes the provided video image and makes recommendations for further actions. The remote assistant significantly improves efficiency of the covert camera detection process and reduces the likelihood of missing potential threats, as he relies on the experience and expert knowledge in the given area. It is possible to use various applications to transmit video information and establish a stable connection, such as Discord or Skype that allow to ensure online transmission of screen recordings.
During the work, the Discord application was used; the image can be enlarged for a clearer examination. In this case, the flare is clearly visible on a large computer screen that makes it possible to detect a video camera with a high probability degree. Moreover, this method increases the likelihood of video camera detection, since several people may participate in the search, one of whom may be an experienced specialist.
In the online mode, it is possible to use neural networks to detect any covert cameras that is based on the efficient use of neural networks for image processing [13]. The successful application of neural networks in recognizing any faces, physical objects, and symbols makes it possible to use this technology to detect covert optics using the flares. The examination pattern allows to exclude a person from the search and perform real time automatic analysis with a remote server using a neural network prepared on the basis of a large data amount, access to which may be paid. The proposed service allows to transfer computational resources from a limited mobile device to a remote server, as well as to compensate for the poor technical specifications of the smartphone’s depth sensor.
C. Room inspection method with a smartphone with a depth sensor to detect covert video cameras using the neural networks
This method provides for the following stages:
Prepare a smartphone for detection of video cameras using a depth sensor by the example of proposals from the previous methods;
Develop and train a neural network with the gathered data using the machine learning and deep learning techniques. The training aim is to teach the neural network to recognize the availability of covert cameras in the images.
Integrate the trained neural network into a mobile application or web service to be used to detect covert cameras in real time. The application must be able to process the video stream from the camera of a smartphone or other device and transmit it to the neural network for analysis.
When using an application to detect covert cameras, the neural network will analyze the video stream, search for the signs of cameras and determine their location in the image.
When the optical device of a covert video camera is detected by a neural network, the application must notify the user while showing the detected camera and providing the opportunity to make decisions (for example, take a photo of the camera, make a report, or remove it).
The neural network efficiency in detecting any covert cameras will depend on the training data quality and the neural network architecture. The constant data update and additional training of the neural network will increase the detection reliability level and adapt to the new types of covert cameras.
Thus, based on the practical studies performed, three search methods for covert video cameras using a smartphone depth sensor are provided. Each method requires certain technical equipment and various financial costs. In the case of using data processing technologies with the neural networks, some organizational and technical resources are required to provide remote access, allocate the cloud resources for data processing, and perform practical training work in relation to the neural network. The covert optics detection method using the neural networks seems to be the most promising one due to the rapid development of pattern recognition technologies in the computer vision systems based on the neural networks.
Conclusion
Practical studies of the personal private space protection against the covert CCTV systems during the private meetings, negotiations and other confidential events by the optical flare recording using a smartphone depth sensor have demonstrated the high efficiency of detecting covert video surveillance systems. The ease of implementation of the examination procedure using the generally available equipment allows to achieve the desired result without any unnecessary costs and efforts.
References
Yarochkin, V. I. Information security: Textbook for universities. “Akad. Project”, Moscow, Russia. (2012).
Horev, A. A. Technical protection of information: studies. a manual for university students. In 3 volumes. Volume 1. Technical channels of information leakage, NPC “Analytics”, Moscow, Russia. (2008).
Horev, A. A. “Means of hidden video surveillance and shooting (based on foreign press materials)”, Special’naya tekhnika, no. 3, pp. 2–23, (2010).
Alferov, V. Yu., Fedyunin, A. E., Peretyatko, N. M. Special equipment of internal affairs bodies. The use of operational surveillance tools in the fight against crime: a textbook. SSEI REU im. G. V. Plekhanova, Saratov, Russia. (2012).
Horev, A. A. “Means of detecting hidden video surveillance systems”, Special’naya tekhnika, no. 6, pp. 53–61. (2015)
Akhmanov S. A., Nikitin S. Yu. Physical optics. “Publishing House of Moscow State University”, “Nauka”, Moscow, Russia. (2004).
Ulysse Delabre Smartphones: scientific experiments with a smartphone / Translated from the French by P. Y. Sergeeva; edited by V. I. Petrovichev. / “DMK Press”, Moscow, Russia. (2021).
Dal Mutto C., Zanuttigh P. and Cortelazzo G. M. Time-of-Flight Cameras and Microsoft KinectTM. A user perspective on technology and applications. Published in “Springer Briefs in Electrical and Computer Engineering”. (2012) DOI:10.1007/978‑1‑4614‑3807‑6
Miles Hansard, Seungkyu Lee, Ouk Choi, Radu Horaud Time-of-Flight Cameras: Principles, Methods and Applications. “Springer Science & Business Media”. (2012) DOI: 10.1007/978‑1‑4471‑4658‑2
Sriram Sami, Sean Rui Xiang Tan, Bangjie Sun, and Jun Han LAPD: Hidden Spy Camera Detection using Smartphone Time-of-Flight Sensors. In The 19th ACM Conference on Embedded Networked Sensor Systems (SenSys ’21), November 15–17, 2021, Coimbra, Portugal. ACM, New York, NY, USA, pp 288–301. (2021) DOI: 10.1145/3485730.3485941
New functions of the Galaxy S20 smartphone. [Electronic resource] – Access mode: https://www.samsung.com/ru/support/mobile-devices/check-out-the-new-camera-functions-of-galaxy-s20‑plus-s20‑ultra / (accessed: 2023.05.17)
Description of ToF Viewer. [Electronic resource] – Access mode: https://m.apkpure.com/ru/tof-viewer-night-vision/com.lvonasek.tofviewer (accessed: 2023.05.17)
Simon Haykin Neural networks. A comprehensive foundation. Second Edition. Upper Saddle River, N.J.: Prentice Hall. 1999.
Information about the author
Vladimir V. Grishachev, Cand. of Sci. (Phys.and Math.), associate professor, Russian State University for the Humanities, Moscow, Russia.
ORCID: 0000-0002-7585-7282
Andrey S. Yusupovsky, student, Russian State University for the Humanities, Moscow, Russia.
A. S. Yusupovsky, V. V. Grishachev
Russian State University for the Humanities, Moscow, Russia
The paper presents research materials related to the use of a smartphone video system (depth sensor) to detect the covert video cameras in various premises, including the places of temporary accommodation (hotel rooms, locker rooms, passenger compartments, etc.) where a person suspects an intrusion on his personal security or privacy breach. The depth sensor of a smartphone makes it possible to increase the efficiency of automated search for covert video cameras by optical flare for an untrained person and security specialists. The article formulates proposals for the search procedure for covert video cameras and identifies the main technical specifications of an efficient survey.
Keywords: personal information security, visual information, covert video camera, smartphone depth sensor
Article received: 17.09. 2023
Article accepted: 07.11. 2023
INTRODUCTION
Protection of information related to the personal data in the form of photo/video materials is an important element in ensuring the private space security. Any personal information security threats can occur in various situations. One of such cases is location of a person in his private space, when he must be sure of protection against any covert surveillance, for example, being in a hotel room, in public places with limited visual access, etc. Such threats can be of great importance for the personal information security [1, 2] and result, for example, in the form of unauthorized publishing of obtained photo/video materials on social networks and other web resources.
The issue of information security of a personal private space shall be resolved not only by the state in the form of passage of various laws, provision of special services, etc., but also by the technical capabilities of the citizen. The technical devices that ensure personal information security must be efficient, economically and technically accessible to every person. In the contemporary society, such a device is a smartphone performing a wide range of functions from any communication to entertainment capabilities. The smartphones have a set of sensors that can be used, among other things, to ensure personal information security.
While the security services with powerful technical and legal capabilities are responsible for ensuring the visual information protection at the state level and at the level of large-scale companies, then confidentiality for an individual or a small company shall be maintained exclusively by personal precautions and allowable generally available technical devices. Such restrictions lead to the search for the most efficient and affordable technical solutions that do not burden a person financially or in other ways.
Technical intelligence and visual information protection
The covert video surveillance in the field of technical intelligence can be performed using video cameras for various purposes [2, 3]: from any custom web cameras to the special covert video cameras camouflaged as the ordinary household items. The small-sized pinhole cameras with an extended pupil, a typical lens size of less than a millimeter, and conversion of an optical image into an electronic or digital signal in the silicon CCD matrices and open (broadcasting) or closed (cable) transmission to the recording system are used in the indoor covert CCTV systems at the short distances (up to tens of meters).
The countermeasures against the visual intelligent systems using the covert CCTVs can be divided into two types, related to their functioning:
Covert video surveillance countering measures based on the detection of electromagnetic radiation (EMR) of a running video camera [2, 5]; by searching for the electronic elements by nonlinear location using electromagnetic radiation; as well as neutralizing the system’s electronics by electromagnetic interference or destruction by an electromagnetic pulse. This method has its own limitations related to performance of the activities using special equipment that is not always available.
The Protect K18 detector (Fig. 1) is commercially available. It allows scanning of the air EMR within the range from 1 MHz to 8 GHz and identifying radio communication channels with the external devices with a sensitivity of more than 30 μW and a declared range of 0.1 m to 15 m. The device is small-sized, but does not help to identify non-emitting devices with the internal media recording or transmissions via the cable systems. The application of such systems in the urban environments is limited by the difficult radio environment in the surrounding area.
Covert video surveillance countering measures based on destructive effects on the optical system through significant flash exposure in the optical or infrared region of the EMR spectrum [2]. This method requires either preliminary detection of the surveillance system optical devices, or isotropic flash exposure of the entire space that needs the comprehensive provision of special equipment. The most efficient method for identifying covert video surveillance systems is based on the flares developed when optical radiation is reflected from the objective lenses and CCD matrix.
Optical system of a covert video camera and its detection
The basis of any CCTV system is an optical system forming an image of the surrounding space on a CCD matrix [2, 3].
The main optical system for covert video surveillance are the cameras with a pinhole lens (Fig. 2), i. e. a lens with a pupil (1) remote at a distance l with a diameter d (less than 1 mm) and typical dimensions of several mm, determined by the focal length f of the lens (2). The input optical radiation is focused on a CCD matrix (3) that converts the optical signal into an electrical signal (4). The incoming light flux at the viewing angle ϕ is subject to the Fresnel reflections on the focusing lenses of the objective and matrix (5) that are perceived as the optical flares, i. e. bright light spots on a uniform background in the visible and other regions of the optical spectrum. Such flare detection in the visible spectrum is quite difficult.
As an example, it is possible to use a laptop camera under the protective glass (Fig. 3). In this case, the video camera can be detected in the visible light by the design features of protective shutter and reflective coating around the camera pin-hole (dark dot in the light outline on the right figure). In the case of oblique shooting, the pin-hole is not visible, but in the case of ordinary light incidence it is difficult to distinguish against a dark background for the human eye, even it is located at the close distance (a laptop). To reduce the Fresnel losses related to the reflections, the camera’s optical elements are coated by the antireflection films that reduce reflection by 3–4 times in comparison to the standard value of 4% [6]. In the case of ordinary light, such surfaces form weak reflections and are therefore difficult to detect with the eye. The antireflection coating relates to the central region of the visible spectrum and, to a lesser extent, visible red light and, even to a lesser extent, invisible infrared light. It is possible to increase observability by moving from the visible region to the visible red spectrum part, by illuminating with intense red light and observing the area with a video camera through a red color filter.
The simplest detection system for video camera lenses includes the devices such as the Protect K18 detector (Fig. 1) that has several LED red light sources. When observing a surface irradiated with the red light through a red color filter in the detector’s center, the optical surfaces form flares visible to the eye (if available). This method depends on the experience of an examiner and requires meticulous long-term work when examining the large surfaces. Any professional cover video camera detectors apply more powerful monochrome LED light sources with a small viewing angle of several degrees and an electronic light detection system that make it possible to register reflective objects at the distances of tens of meters.
Transition to the invisible infrared portion of the spectrum increases the Fresnel reflection that is not blocked by antireflection coatings. The brightness of flares, invisible to the eye, is increased, therefore, its observation requires application of the infrared radiation receivers.
Smartphone video system
and depth sensor
The up-to-date wearable devices, such as smartphones, can provide an affordable and efficient solution to protect individuals against any visual reconnaissance [7]. A smartphone is a generally accessible device with a wide range of functionality. Firstly, it is a communication device that allows perform local communication using Bluetooth technology, Wi-Fi and long-distance communication using 4G/5G technology, etc. Secondly, the smartphones include a large number of sensors from the sensors converting audio and video information to the physical field sensors. All sensors can be built-in or external with a wired or wireless connection. Any smartphone can be used as a computing device to measure any physical quantities using its own or connected sensors, the capabilities of which are limited only by the power supplies and computational resources [7].
One of the advanced smartphone systems is its video system for visual information recording and display. For example, Samsung Galaxy A22 (Fig. 4) has five video cameras for various purposes: a front-facing camera for selfies, a macro (close-up) camera, an ultra-wide-angle camera for shooting the close-up large objects or panoramic shooting, a main camera for the high-quality portrait photographs, a depth camera for sharpening the frame and other purposes. In addition to the cameras noted, there are some built-in cameras for night photography, black and white cameras and others that can be of external nature, such as the thermal imaging cameras.
In smartphones, 3D images are created using the stereoscopic vision systems based on two cameras (stereo vision); structural light projection by infrared radiation in the form of a patterned set of points and lines distributed in the space, the curvature of which determines the three-dimensional object structure (structured light). Recently, the iPad Pro 2020 mobile devices are provided with the lidar systems. The depth sensors are the most efficient option for creating a three-dimensional image (ToF sensors, time-of-flight cameras, ToF cameras, depth cameras) to be used to measure distance, recognize gestures and faces, develop 3D object models, proximity sensor and for other purposes [8, 9, 10].
The depth sensor (ToF-camera) is based on measuring the propagation time of a light pulse from an infrared laser to a reflective obstacle and back to a detector in the form of an IR camera (Fig. 5). The IR laser (2) of the depth sensor (1) generates the modulated optical radiation directed at the object under study being synchronized with the IR camera operation process (3). The optical radiation (4, 5) reflected from various objects (6, 7, 8) is recorded by an IR camera with a time delay. The duration of exposure to the laser radiation is several nanoseconds that makes it possible to record the phase shift of reflected radiation with the millimetric accuracy. When processing the delay time (time of flight), a 3D image of the probed space is generated by isolating the areas with the same delay time to be highlighted in brightness and color. Various shades of red light are usually used: the closer areas are brighter, and the more distant ones are dimmer.
The 3D image generation by the depth sensor camera depends on the IR laser intensity and the IR camera sensitivity; the higher the radiation intensity, the greater the reflected signal intensity and the less error of the distance to be determined. As the distance of the object under study is increased, intensity of the reflected IR radiation decreases leading to the enhanced measurement error. On the other part, as the distance to the object is decreased, the time of flight is reduced and the distance measurement error is also increased. Consequently, each depth sensor has its own region of space in terms of a distance at which it operates most efficiently.
The efficiency of a response generation from objects in the form of holes with a small reflective diameter and shallow depth, such as a covert camera with a pinhole lens, depends on the field of view (viewing angle) of the hole (lens). Only when the laser radiation enters the field of view of the lens, the response is formed that imposes another limitation for observing such objects, namely the viewing angle of the observed covert camera hole and the source of laser radiation. Displacement of the depth sensor along the surface under study must be within the viewing angles of the lens and laser.
Any proposals for using a depth sensor to detect a covert camera are given in some papers [10]. In particular, it is proposed to develop software for installation on a smartphone with a depth sensor to detect any images obtained from the small holes in the form of flares, i. e. significant small reflections, clearly visible against a generally uniform background. Such papers do not discuss connection with the technical parameters of a smartphone, so our paper focuses on capabilities of the smartphone’s depth sensor and possible practical inspection of premises for the availability of covert cameras.
Practical simulation of video camera detection using a smartphone’s depth sensor
The practical studies of detecting covert video cameras were performed using a Samsung Galaxy S 20+ smartphone with a built-in DeethVision ToF-camera (Fig. 4) [11]. The depth sensor operates in the ambient imaging mode using the open source ToF Viewer Android application [12]. It shows the resulting video image on the smartphone display in real time. The DeethVision camera functions in the invisible IR range with its own source that accounts for the maximum sensitivity of the silicon CCD matrix. The generated depth map is perfectly visible in absolute darkness.
The practical studies consisted of demonstration of the possible camera detection by the flare of optics, determination of technical specifications of the depth sensor of a given smartphone, for which a convert video camera model was prepared. On the basis of experimental studies performed, several proposals were made for practical implementation of the covert CCTV detection process.
Demonstration of the ability to detect video cameras using a smartphone’s depth sensor
Verification of the possible camera detection using a smartphone’s depth sensor was carried out by scanning of the external web camera (Fig. 6) and a laptop web camera (Fig. 7). The photographs given were taken with a conventional camera (on the left) and a camera with a depth sensor (on the right) at a normal location at various distances (farther and closer). As one can see from the photographs in normal light, even when placed closely, it is not possible to distinguish the webcam lens of both the external and built-in laptop cameras. It is possible to determine the camera existence by the light-colored rings around the lens entrance hole typical for the web cameras. These differences are more clearly visible during the eyeball inspection when the viewing angle is changed. When approaching the external webcam, it is possible to distinguish the pinhole lens. When using a camera with a depth sensor, the pinhole lens is clearly visible when viewed closely and is difficult to see when moving away from the object. Thus, it is possible to confirm a covert video camera detection process based on the flares from a pinhole lens using a camera with a depth sensor.
Test model of a covert video camera
A WI-FI camera was used as a control video camera simulating the covert video surveillance system (Fig. 8). In order to increase the external covertness attributes by reducing the physical dimensions, the optical unit (lens with a CCD matrix in the housing) was removed from the camera. This activity reduced the camera dimensions while completely disrupting its functionality. The physical parameters of the optical unit were determined by the hole size for the input optics that was about 3 mm, with a total diameter of the cylindrical part of the camera optical unit of 20 mm and a thickness of 24 mm.
The camera was camouflaged to match the external background using a black sheet of paper, where a hole was made for the lens (Fig. 9). The black paper background faded with the camera’s black color that simulated the camera placement secrecy. During the simulation process, the task of preserving the shooting functions and completely concealing the camera placement was not set, i. e. only the conditional simulation of functioning and placement of a covert video camera was provided.
All subsequent experimental studies were performed with the described simulated covert camera in the form of a lens inserted into a hole in a black sheet of paper (Fig. 8 and Fig. 9).
Experimental determination of technical capabilities of the depth sensor in a Samsung Galaxy S 20+ smartphone
The test objective was to determine the technical parameters for a simulated covert video camera detection using a camera with a smartphone depth sensor, such as the optimal distance to the object and the maximum viewing angle at which optical flare was observed most clearly.
The optical circuit of the experimental studies is shown in Fig. 10. The covert camera model (1) against the black screen (2) was placed at a certain distance l opposite to the smartphone camera (3) and the optical flare was recorded. The smartphone approached the object at a minimum distance and then moved away until the flare disappeared completely at a distance lmax. As a result of this test, the maximum observation distance lmax of a covert camera based on the flare was recorded with the smartphone in normal position. Further, the camera was placed at a distance lmax / 2 being half the maximum value for optimal observation of flares. At this distance, the smartphone moved parallel to the screen at a distance lplus, when the flares discontinued according to which the camera viewing angle ϕt was determined. All results were documented by photographing a model at experimentally determined distances with a conventional camera and a smartphone ToF-camera with visualization using the ToF Viewer application.
At the beginning of the experiment, a distance was selected equal to 40 cm (Fig. 11). With the normal smartphone position and a conventional camera, the covert camera optics were almost not observed. However, the ToF-camera provided its clear observation. The photographs made by the ToF-camera show not only the direct flares, but also the flares obtained when reflected from the table that complicates the task of identifying informative flares using the computer methods.
During the next experiment (Fig. 12), the studies were performed at a distance of 85 cm, when a conventional camera did not allow flares from the covert optics to be seen, and the ToF-camera generated a completely visible image. Moreover, as can be seen in the photographs, additional flares related to the reflections from auxiliary surfaces are not observed, since the side reflected light fluxes of the IR laser do not fall into the field of view of the covert camera and do not generate additional flares as in the previous experiment (Fig. 11). Based on this experiment, it is possible to estimate the limiting angle (about 15° × 2) of the viewing field of a model covert camera as the arctangent of the flare height above the table and the table diameter.
When moving at a distance of 120 cm, it is still possible to observe the flares, but when the smartphone is removed at a distance of more than 140 cm from the model, it leads to a complete loss of flare observation using the ToF-camera (Fig. 13). Thus, the test results confirm that the observation of flares is stopped at a distance lmax = 140 cm. The next stage of the research, namely determination of the camera’s field of view angle, was carried out at a distance of 70 cm from the object with parallel movement of the smartphone relative to the screen. As a result of measurements, the obtained estimates of the field of view limiting angle (Fig. 12) of 15x2 degrees were confirmed.
Experimental comparison of technical capabilities of the depth sensor in a Samsung Galaxy S 20+ smartphone and wiretapping detector Protect K‑18S
As it has been noted above, to search for covert optics, the K‑18 detector can be used. It has a visible red light source and a red optical filter to observe the reflected red light through the eyepiece.
A comparison of capabilities of the ToF-camera and the K18 sensor in detecting the lens flares demonstrates the advantages of the first method when inspecting the rooms for the covert camera availability using the lens flares. The results of experimental comparison of the sensors are given in Fig. 14 and fig. 15. As it can be seen from the photographs, the flares recorded by the K‑18 sensor are much weaker than provided by the ToF-camera. In the comparison experiment (Fig. 14 and Fig. 15), the image generated by the K‑18 sensor was recorded by the smartphone camera that approximated the object and had a higher sensitivity to red light than the human eye. In reality, observation must be performed only with the eye through the lens that further complicates the search procedure.
Study results and proposals for detecting covert video cameras
The practical studies performed and the results obtained make it possible to propose the technology and method for conducting inspections of premises and confidence rooms for the availability of covert CCTV systems in order to protect the personal private space. A smartphone with a depth sensor can be used to professionally protect visual information while improving its performance. The detection efficiency of the convert video cameras can be enhanced by increasing power and coherence (monochrome) of the infrared laser radiation that will lead to an increase in the detection range of cameras, due to the higher reflected power, and measurement accuracy of the difference in distances traveled by the laser radiation as a result of phase measurement methods. In practice, to detect any flares from the camera optics, the CCD matrix and smartphone display quality (resolution, brightness, contrast) is of great importance. The given recommendations can significantly increase the detection efficiency of professional covert video cameras with a small input hole (pinhole) of the lens that can be less than 1 mm while developing low intensities of reflected radiation.
Based on the research results, it is possible to identify the necessary inspection conditions for searching for video cameras:
- availability of a smartphone with a depth sensor and visualization software;
- practical evaluation of the sensor to detect cameras by the flares based on the available ones, such as a laptop webcam or other cameras;
- experience in obtaining the basic detection skills;
- determination of sensor parameters for detecting the covert optics.
Based on these requirements, several real time room inspection methods can be proposed.
A. Room inspection method (including confidence rooms) using a smartphone with a depth sensor to detect covert video cameras
Setup/preparation/calibration. While using standard available video cameras, configure a smartphone with a depth sensor for detection: determine the type of flares, establish the maximum observation length lmax and distance lplus;
Conduct an inspection of the room walls with image recording, placing the smartphone camera at the optimal distance lmax/2 and moving it with swinging parallel to the wall;
Visually identify the hotspots suspected for the availability of video cameras based on the flares occurred;
Display the video recording on a large monitor screen in the slow-motion mode: visually re-identify the hotspots suspected for the availability of video cameras based on the flares occurred;
Separately examine the hotspots at a closer distance with slow movement of the smartphone’s video camera;
Draw conclusions about the availability of covert video cameras.
When the person examining the premises does not have significant experience or does not have the physical capabilities for an effective survey, such as, for example, poor vision, poor attention to details, etc., he can involve any remote assistants in the examination or conduct the examination in the online mode using the smartphone’s communication capabilities. In this case, the method will be changed while maintaining all the basic principles. However, the process efficiency and velocity can be increased.
B. Room inspection method by a smartphone with a depth sensor to detect covert video cameras in the online mode with an assistant
Prepare a smartphone for detection of video cameras using a depth sensor by the example of proposals from the previous methods;
Establish an online connection with a remote assistant for real time transmission of video images to a remote large monitor;
Monitor the premises with the constant voice communication with an assistant.
At the best case, the assistant is a specialist or expert in the field of covert camera detection who is located at a distance and provides the real time support and advice. He helps during the examination for availability of covert video cameras, analyzes the provided video image and makes recommendations for further actions. The remote assistant significantly improves efficiency of the covert camera detection process and reduces the likelihood of missing potential threats, as he relies on the experience and expert knowledge in the given area. It is possible to use various applications to transmit video information and establish a stable connection, such as Discord or Skype that allow to ensure online transmission of screen recordings.
During the work, the Discord application was used; the image can be enlarged for a clearer examination. In this case, the flare is clearly visible on a large computer screen that makes it possible to detect a video camera with a high probability degree. Moreover, this method increases the likelihood of video camera detection, since several people may participate in the search, one of whom may be an experienced specialist.
In the online mode, it is possible to use neural networks to detect any covert cameras that is based on the efficient use of neural networks for image processing [13]. The successful application of neural networks in recognizing any faces, physical objects, and symbols makes it possible to use this technology to detect covert optics using the flares. The examination pattern allows to exclude a person from the search and perform real time automatic analysis with a remote server using a neural network prepared on the basis of a large data amount, access to which may be paid. The proposed service allows to transfer computational resources from a limited mobile device to a remote server, as well as to compensate for the poor technical specifications of the smartphone’s depth sensor.
C. Room inspection method with a smartphone with a depth sensor to detect covert video cameras using the neural networks
This method provides for the following stages:
Prepare a smartphone for detection of video cameras using a depth sensor by the example of proposals from the previous methods;
Develop and train a neural network with the gathered data using the machine learning and deep learning techniques. The training aim is to teach the neural network to recognize the availability of covert cameras in the images.
Integrate the trained neural network into a mobile application or web service to be used to detect covert cameras in real time. The application must be able to process the video stream from the camera of a smartphone or other device and transmit it to the neural network for analysis.
When using an application to detect covert cameras, the neural network will analyze the video stream, search for the signs of cameras and determine their location in the image.
When the optical device of a covert video camera is detected by a neural network, the application must notify the user while showing the detected camera and providing the opportunity to make decisions (for example, take a photo of the camera, make a report, or remove it).
The neural network efficiency in detecting any covert cameras will depend on the training data quality and the neural network architecture. The constant data update and additional training of the neural network will increase the detection reliability level and adapt to the new types of covert cameras.
Thus, based on the practical studies performed, three search methods for covert video cameras using a smartphone depth sensor are provided. Each method requires certain technical equipment and various financial costs. In the case of using data processing technologies with the neural networks, some organizational and technical resources are required to provide remote access, allocate the cloud resources for data processing, and perform practical training work in relation to the neural network. The covert optics detection method using the neural networks seems to be the most promising one due to the rapid development of pattern recognition technologies in the computer vision systems based on the neural networks.
Conclusion
Practical studies of the personal private space protection against the covert CCTV systems during the private meetings, negotiations and other confidential events by the optical flare recording using a smartphone depth sensor have demonstrated the high efficiency of detecting covert video surveillance systems. The ease of implementation of the examination procedure using the generally available equipment allows to achieve the desired result without any unnecessary costs and efforts.
References
Yarochkin, V. I. Information security: Textbook for universities. “Akad. Project”, Moscow, Russia. (2012).
Horev, A. A. Technical protection of information: studies. a manual for university students. In 3 volumes. Volume 1. Technical channels of information leakage, NPC “Analytics”, Moscow, Russia. (2008).
Horev, A. A. “Means of hidden video surveillance and shooting (based on foreign press materials)”, Special’naya tekhnika, no. 3, pp. 2–23, (2010).
Alferov, V. Yu., Fedyunin, A. E., Peretyatko, N. M. Special equipment of internal affairs bodies. The use of operational surveillance tools in the fight against crime: a textbook. SSEI REU im. G. V. Plekhanova, Saratov, Russia. (2012).
Horev, A. A. “Means of detecting hidden video surveillance systems”, Special’naya tekhnika, no. 6, pp. 53–61. (2015)
Akhmanov S. A., Nikitin S. Yu. Physical optics. “Publishing House of Moscow State University”, “Nauka”, Moscow, Russia. (2004).
Ulysse Delabre Smartphones: scientific experiments with a smartphone / Translated from the French by P. Y. Sergeeva; edited by V. I. Petrovichev. / “DMK Press”, Moscow, Russia. (2021).
Dal Mutto C., Zanuttigh P. and Cortelazzo G. M. Time-of-Flight Cameras and Microsoft KinectTM. A user perspective on technology and applications. Published in “Springer Briefs in Electrical and Computer Engineering”. (2012) DOI:10.1007/978‑1‑4614‑3807‑6
Miles Hansard, Seungkyu Lee, Ouk Choi, Radu Horaud Time-of-Flight Cameras: Principles, Methods and Applications. “Springer Science & Business Media”. (2012) DOI: 10.1007/978‑1‑4471‑4658‑2
Sriram Sami, Sean Rui Xiang Tan, Bangjie Sun, and Jun Han LAPD: Hidden Spy Camera Detection using Smartphone Time-of-Flight Sensors. In The 19th ACM Conference on Embedded Networked Sensor Systems (SenSys ’21), November 15–17, 2021, Coimbra, Portugal. ACM, New York, NY, USA, pp 288–301. (2021) DOI: 10.1145/3485730.3485941
New functions of the Galaxy S20 smartphone. [Electronic resource] – Access mode: https://www.samsung.com/ru/support/mobile-devices/check-out-the-new-camera-functions-of-galaxy-s20‑plus-s20‑ultra / (accessed: 2023.05.17)
Description of ToF Viewer. [Electronic resource] – Access mode: https://m.apkpure.com/ru/tof-viewer-night-vision/com.lvonasek.tofviewer (accessed: 2023.05.17)
Simon Haykin Neural networks. A comprehensive foundation. Second Edition. Upper Saddle River, N.J.: Prentice Hall. 1999.
Information about the author
Vladimir V. Grishachev, Cand. of Sci. (Phys.and Math.), associate professor, Russian State University for the Humanities, Moscow, Russia.
ORCID: 0000-0002-7585-7282
Andrey S. Yusupovsky, student, Russian State University for the Humanities, Moscow, Russia.
Readers feedback
