Camera and method for facilitating mounting of the camera

Abstract

A video camera and a method for facilitating installation of the video camera are provided. The method comprises capturing a video stream of a scene (102) including an object (104) when the video camera is installed with a first FOV. A size of at least one portion of the object in the video stream is estimated. A first edge-filtered version of the video stream is provided to ensure privacy of the object as long as the estimated size is above a first threshold and below a second threshold. In response to the estimated size being above the second threshold indicative of a risk of privacy breach, an enhanced privacy mode of operation is entered, including generating an alert to prompt an installer to reinstall the video camera with a second FOV different from the first FOV, and disabling the provision of the first edge-filtered version of the video stream.

Description

Technical Field

[0001] The present invention relates to a camera and a method for facilitating the installation of the camera. Background Technology

[0002] When cameras, such as surveillance cameras, are installed to capture images of a scene, national privacy regulations may require that the privacy of one or more objects present in the scene be guaranteed. Privacy is often achieved by distorting the captured image to prevent object identification or recognition. For example, distorting facial recognition in the captured image can prevent vehicle registration number identification. If privacy cannot be guaranteed, the installation of cameras is not permitted.

[0003] Methods for distorting captured images to prevent object recognition and identification are known. For example, US10755388B2 describes how to protect personal privacy in captured image data by controlling privacy masking of the image, wherein the degree to which privacy areas to be applied to the image are blurred depends on a threshold distance from the image capturing device and the spatial resolution of the scene content in the image at that distance. However, US10755388B2 does not disclose how to facilitate camera installation to ensure privacy.

[0004] US2014 / 0362215A1 relates to camera mount calibration. However, in that disclosure, the purpose of camera mount calibration is to optimize face recognition in captured images. The average size of the face is measured in different sub-regions and used as a parameter to trigger changes in camera position, such as moving the camera further away from the scene or zooming in.

[0005] Therefore, there is room for improvement in order to provide a camera that can be easily installed in accordance with national privacy regulations. Summary of the Invention

[0006] Therefore, in view of the above, the object of the present invention is to mitigate the disadvantages of the prior art and to provide a technique for facilitating camera installation, and in particular a technique for ensuring privacy when capturing video streams.

[0007] According to a first aspect of the invention, the above objective is achieved by a method performed by a camera to facilitate camera installation. The method includes: capturing a video stream of a scene, the scene comprising one or more objects, when the camera is installed with a first field of view. The method further includes: estimating the size of the video stream of at least a portion of the one or more objects, and providing a first edge-filtered version of the video stream to ensure the privacy of the one or more objects depicted in the video stream, provided that the estimated size of the video stream of at least a portion of the one or more objects is above a first threshold and below a second threshold. Furthermore, the method includes: in response to the estimated size of the video stream of at least a portion of the one or more objects being above a second threshold indicating a risk of privacy breach of at least one object, entering an enhanced security and privacy operating mode, including: generating an alarm to prompt the installer to reinstall the camera with a second field of view different from the first field of view, and prohibiting the provision of the first edge-filtered version of the video stream.

[0008] A first edge-filtered version of the video stream is provided if the estimated size of at least one portion of the video stream of one or more objects is above a first threshold and below a second threshold. The privacy of one or more objects depicted in the video stream is ensured by anonymizing the one or more objects provided by the first edge-filtered version of the video stream.

[0009] Furthermore, the privacy of one or more objects is ensured by prohibiting the provision of a first edge-filtered version of the video stream when the estimated size of at least a portion of the video stream of one or more objects exceeds a second threshold indicating a risk of privacy breach for at least one object.

[0010] An alarm is generated to prompt the installer to reinstall the camera when the estimated size of at least a portion of the video stream of one or more objects exceeds a second threshold. This informs the installer that the camera needs to be reinstalled with a second field of view, different from the first, to ensure privacy in future captures of the video stream. Therefore, the second field of view should be selected such that it will provide privacy for one or more objects captured in the video stream.

[0011] When used in this disclosure, the term "mount" refers to the installation of a camera at a location (e.g., at a physical location, at a scene) and / or the configuration of a camera used to capture a video stream of the scene in a certain field of view. Thus, "mount" can refer to the geographical location where the camera is mounted and / or the camera settings (such as a pan, tilt, and / or zoom setting) used to capture the video stream.

[0012] As used herein, the term field of view refers to the portion of the scene visible through the camera at a specific location and orientation, and with a specific zoom setting. The specific location is given by the mounting position, and the orientation is given by the camera's pan and / or tilt settings. Therefore, it should be understood that the field of view can depend on one or more different camera parameters. For example, the field of view can depend on the camera's mounting position (such as its height above the ground, the camera's zoom setting, the camera's pan setting, and / or the camera's tilt setting).

[0013] The estimated size of at least one portion of a video stream of one or more objects can be determined as the number of pixels in at least one portion formed along at least one direction of the video stream. For example, the estimated size of at least one portion can be given as the number of pixels in the vertical (height) and / or horizontal (width) directions of an image frame in the video stream. A pixel (sometimes referred to as a picture element) is the smallest single element in an image frame. Depending on the spatial resolution of the image frame, an image frame can include 1920 × 1080 (width × height) pixels, 2560 × 1440 pixels, 2560 × 1920 pixels, or 3840 × 2160 pixels, to name just a few examples.

[0014] The first threshold can be zero pixels, provided that a first edge-filtered version of the video stream should be provided for all estimated sizes below a second threshold. However, if the estimated size of the video stream for at least a portion of one or more objects is small enough, the captured video stream can be provided as is without risk of disclosing the privacy of one or more objects. What should be considered a sufficiently small estimated size depends on the situation, but a higher threshold should be set to provide a higher threshold when the video stream can be provided as is without disclosing the privacy of one or more objects depicted in the video stream. In this case, a first edge-filtered version of the video stream is provided when the estimated size of the video stream for at least a portion of one or more objects is higher than the first threshold but lower than the second threshold.

[0015] When a first edge-filtered version of the captured video stream should be provided, a second threshold should be set to provide a higher threshold. Above this second threshold, if the first edge-filtered version of the video stream is provided, the privacy of one or more objects is at risk of being compromised, and therefore, the provision of the first edge-filtered version of the video stream is prohibited.

[0016] The first and second thresholds can be determined based on camera parameters and installation location. For example, the thresholds can be determined based on the camera's optics (such as lens type) and the camera's resolution and field of view. Additionally or alternatively, the thresholds can also be determined based on an edge filter used to provide an edge-filtered version of the captured video stream and the kernel size based on the edge filter. The kernel is a small matrix used in edge detection when performing convolution between the kernel and image frames. Furthermore, the first and second thresholds determined for a camera installation in one country may differ from those determined for a camera installation in another country due to differences in that country's privacy regulations. The first and second thresholds can be determined by performing tests or based on the first and second thresholds used for previous camera installations. The first and second thresholds can also be determined using machine learning techniques (e.g., by a neural network configured to detect and / or identify one or more objects in the video stream).

[0017] It should be understood that a video stream includes at least one image frame of a scene. Therefore, a video stream may include one or more image frames (e.g., multiple temporally consecutive image frames).

[0018] A scene can be any three-dimensional physical space whose size and shape are defined by the field of view of the camera recording the scene.

[0019] When used herein, the term "object" means any kind of object that can reveal private information about the object itself, about the person who owns the object, or about the person associated with the object. Thus, an object can be a person or a part thereof (such as a face or body part), a group of people, a vehicle or a part thereof (such as a vehicle license plate), and so on, to name a few.

[0020] An edge-filtered version of a video stream is a version of the video stream in which edge points have been enhanced and non-edge points have been suppressed. Edge points correspond to points in the image frames of the video stream where the image brightness changes abruptly or, more formally, where there is a discontinuity. Correspondingly, non-edge points correspond to pixels in the image frames where the image brightness does not change abruptly or, more formally, where there is no discontinuity. In the edge-filtered version of the video stream, edge points are enhanced and are typically represented as a set of connected curves indicating the boundaries of objects, the boundaries of surface markers, and / or a set of curves corresponding to discontinuities in surface orientation. Because non-edge points are suppressed in the edge-filtered version of the video stream, points within the boundaries enhanced by edge points are also suppressed. For example, if a facial contour is enhanced in the edge-filtered version of the video stream, surface textures within the facial contour will be suppressed, resulting in a degree of facial anonymization. The degree of anonymization depends on the degree of enhancement and suppression in the edge-filtered version of the video stream. However, it should be understood that the degree of anonymization also depends on the spatial resolution of the video stream.

[0021] In some variations of the method, enhanced security and privacy operating modes include providing a second version of the video stream that further ensures the privacy of one or more objects.

[0022] By providing a second version of the video stream that further ensures the privacy of one or more objects, the privacy of one or more objects can be guaranteed when the first edge-filtered version of the video stream fails to guarantee their privacy and is therefore disabled. Thus, a second version of the video stream can be provided to users for monitoring or storage without disclosing any privacy-sensitive information.

[0023] In some variations of the method, providing a second version of the video stream includes generating a second version of the video stream by applying a privacy mask to at least one portion of the video stream to ensure the privacy of one or more objects depicted in the video stream.

[0024] This is the first alternative to providing a second version of video streaming to guarantee the privacy of one or more objects.

[0025] When used in this disclosure, the term privacy masking should be understood as a privacy masking feature used to protect personal privacy by employing a masked area to hide portions of an image frame from view. Privacy masks can be static or dynamic. Static privacy masks typically remain in place until the operator decides to move or remove them. Dynamic privacy masks can change over time, and the operator can also define when privacy masks should be applied. For example, the operator can define a dynamic privacy masking such that if a face is detected within the masked area, the face will be masked, but in other cases, the masking will not be applied to that area.

[0026] Privacy masking is typically applied to images as an overlay. Privacy masks can be polygonal in shape, but can also be other shapes that more closely resemble the shape of the area to be masked. Some privacy masks take the form of opaque regions (e.g., uniform black or blue areas), while others take the form of blurring, where image data is "smeared" or pixelated over the privacy-masked area. Pixelation is the process where a single pixel value (typically the average of several or all pixels within a coding unit) replaces all pixels in that coding unit. Pixelation can be achieved by reducing the image frame by a factor, for example, between 8 and 64, and then enlarging the image frame using a trivial scaling algorithm such as the nearest neighbor algorithm. While blurring is generally considered preferable to pixelation for privacy masking because it allows for smoother edges and therefore looks better, it is more difficult to achieve and requires more computational resources compared to methods currently feasible in hardware.

[0027] In a video stream, privacy masking is applied to each image frame of the video stream to hide the privacy of one or more objects depicted in the video stream. For example, privacy masking can be applied to the face or body parts of a person and thereby hide these parts from being seen and identified. Privacy masking can be applied to at least one portion of the video stream by defining an area that covers one or more objects depicted in the video stream and representing that area by using solid color, blurring, or pixelation as described above. The area can be defined, for example, as the outline of one or more objects. However, the area can have other shapes suitable for covering at least one portion of the video stream, such as circles, rectangles, or ellipses.

[0028] In some variations of the method, providing a first edge-filtered version of the video stream includes generating the first edge-filtered version of the video stream by applying an edge filter to the video stream to enhance edge points in the video stream and suppress non-edge points in the video stream, wherein the pixel values ​​of the video stream are represented using a first number of quantization levels.

[0029] As described above, by generating a first edge-filtered version of the video stream, the privacy of one or more objects depicted in the video stream is ensured by anonymizing one or more objects provided by the first edge-filtered version of the video stream.

[0030] An edge filter is a filter that enhances edge points and suppresses non-edge points. Therefore, by applying an edge filter to a video stream, edge points are enhanced while non-edge points are suppressed. For example, an edge filter (sometimes also called an edge detection filter) can be a Sobel filter, a Canny filter, a Gaussian difference (DoG) filter, a Laplacian filter, a Prewitt filter, or a Scharr filter, to name just a few examples. To suit different use cases, the output of the edge filter (i.e., the edge-filtered version of the video stream) can be modified using one or more operations, such as absolute value operations that do not distinguish between positive and negative gradients, gain operators that adjust the intensity, offset operators that make edge points black instead of white, and / or noise reduction operators that remove unwanted sparse black edge pixels instead of being part of true edges or other structures in the image.

[0031] As mentioned above, pixels in a video stream are represented using a first number of quantization levels. For example, a pixel can be represented using 8 bits, corresponding to 256 quantization levels. Therefore, each pixel can take a number between 0 and 255 to represent one of 256 possible colors. After applying an edge filter to the video stream, the resulting edge-filtered version of the video stream typically consists of pixels with grayscale values, meaning they only represent the amount of light; that is, each pixel only represents intensity information. In the case of 8-bit representation, each pixel in the edge-filtered version of the video stream will be able to use grayscale to represent an intensity level between 0 and 255. Therefore, if 0 is represented by white and 255 by black, there are 254 grayscale levels between them.

[0032] In some variations of the method, providing a second version of the video stream includes generating a second version of the video stream as a second edge-filtered version of the video stream by reducing the spatial resolution of the video stream before applying an edge filter to the video stream so as to reduce the size of the video stream of at least one portion of one or more objects to below a second threshold.

[0033] This is a second alternative way to provide a second version of video streaming to guarantee the privacy of one or more objects. For some use cases, this can be an effective way to guarantee the privacy of one or more objects.

[0034] When used herein, the term spatial resolution refers to the number of pixels used in the construction of a video stream (e.g., in the construction of one or more image frames of the video stream). By reducing the spatial resolution of a video stream, the number of pixels in one or more image frames that make up the video stream is reduced. This is sometimes referred to as downsampling or reduction. Conversely, by increasing the spatial resolution of a video stream, i.e., by upsampling / enlarging the video stream, the number of pixels in one or more image frames that make up the video stream is increased. It should be understood that the size of each pixel remains the same, resulting in a smaller image size when the spatial resolution is reduced and an larger image size when the spatial resolution is increased. Thus, if provided to the user regardless of privacy level, a video stream (e.g., the original captured video stream), a first version of the video stream, and a second version of the video stream will all have the same pixel size. However, it should be understood that after video stream downsampling, video stream upsampling can be performed to maintain the same number of pixels in the video stream but with less sensitive information because image details are lost during downsampling.

[0035] By reducing the spatial resolution of the video stream so that the size of the video stream for at least a portion of one or more objects is below a second threshold, the privacy of one or more objects obtained in the second edge-filtered version of the video stream is guaranteed.

[0036] In some variations of the method, providing a second version of the video stream includes generating a second version of the video stream as a second edge-filtered version of the video stream by reducing the number of quantization levels used to represent the pixel values ​​of the video stream from a first number of quantization levels to a second number of quantization levels before applying the first edge filter to the video stream.

[0037] This is a third alternative method for providing a second version of the video stream to guarantee the privacy of one or more objects. It should be understood that pixel values ​​in this second edge-filtered version of the video stream will be represented using a second number of quantization levels.

[0038] By reducing the number of quantization levels in the video stream from a first number of quantization levels to a lower second number of quantization levels, the graininess in the second edge-filtered version of the video stream increases because the number of intensities represented by each quantization level is larger in the second number of quantization levels compared to the number of intensities represented by each quantization level in the first number of quantization levels. This greater graininess ensures the privacy of one or more objects depicted in the second edge-filtered version of the video stream.

[0039] For example, if the first number of quantization levels is 256, represented by 8 bits as described above, the second number of quantization levels can be 4 (represented by 2 bits). In this case, the 256 quantization levels of the first number of quantization levels, each representing an intensity, are grouped into 4 groups of 64 quantization levels. The first group can be combined into a first quantization level of the second number of quantization levels (e.g., 0). Therefore, this first quantization level of the second number of quantization levels (e.g., 0) will represent each of the 64 different intensities represented by quantization levels 0 to 63 of the first number of quantization levels. Therefore, in the second edge-filtered version of the video stream, no difference between the 64 different intensities represented by quantization levels 0 to 63 of the first number of quantization levels will be visible. Accordingly, the second group can be combined into a second quantization level of the second number of quantization levels (e.g., 1), and this second quantization level (e.g., 1) will represent each of the 64 different intensities represented by quantization levels 64 to 127 of the first number of quantization levels. Therefore, in the second edge-filtered version of the video stream, no difference will be visible between the 64 different intensities represented by quantization levels 64 to 127 of the first quantization level number in the first edge-filtered version of the video stream. The same applies to the third and fourth groups.

[0040] In some variations of the method, providing a second version of the video stream includes: generating a second version of the video stream as a second edge-filtered version of the video stream by reducing the spatial resolution of a first edge-filtered version of the video stream to obtain a second edge-filtered version of the video stream, wherein the size of at least one portion of one or more objects is below a second threshold.

[0041] This is the second version of providing video streaming, and the fourth alternative to guarantee the privacy of one or more objects.

[0042] A second edge-filtered version of the video stream is obtained by reducing the spatial resolution of a first edge-filtered version of the video stream, wherein the size of at least one portion of one or more objects is below a second threshold, thus ensuring the privacy of one or more objects.

[0043] In some variations of the method, providing a second version of the video stream includes generating a second version of the video stream as a second edge-filtered version of the video stream by representing edge points with edge intensities below an edge intensity threshold in a first edge-filtered version of the video stream as non-edge points in a second edge-filtered version of the video stream.

[0044] This is a fifth alternative method for providing a second version of the video stream to guarantee the privacy of one or more objects. By representing edge points with edge strength below a threshold in the first edge-filtered version of the video stream as non-edge points in the second edge-filtered version, the contours of one or more objects are further enhanced while surface textures within the contours of one or more objects are further suppressed. This results in a higher degree of anonymization for one or more objects, and thus ensures their privacy. The term "edge strength" refers to a measure of the edge and how image brightness changes abruptly at the edge.

[0045] In some variations of the method, providing a first edge-filtered version of the video stream includes transmitting the first edge-filtered version of the video stream to the receiver; and disabling the provision of the first edge-filtered version of the video stream includes disabling the transmission of the first edge-filtered version of the video stream to the receiver.

[0046] The receiver can be any device that is directly or indirectly connected to the camera via a network, operatively or wirelessly. For example, the receiver can be a storage device, a user terminal, a server, or a cloud device.

[0047] In some variations of the method, the provision of the first edge-filtered version of the video stream blocking includes: blocking the camera from capturing the video stream. Blocking the camera from capturing the video stream means that the capture of the video stream is terminated. Thus, if provided, there is no captured video stream that could potentially expose the privacy of one or more objects.

[0048] In some variations of the method, providing a first edge-filtered version of the video stream prohibition includes: prohibiting the camera's image sensor from transmitting image information used to generate the captured video stream.

[0049] By prohibiting the image sensor from transmitting image frames used to generate the captured video stream, there is no image information to be processed, and therefore, if provided, there is no captured video stream that could compromise privacy.

[0050] In some variations of the method, the method further includes providing the captured video stream when the estimated size of the video stream of at least one portion of one or more objects is below a first threshold.

[0051] Therefore, when the size of the video stream of at least one portion of one or more objects is below a first threshold, it is concluded that the captured video stream will not compromise privacy if provided as is. Therefore, the captured video stream is provided when the size of the video stream of at least one portion of one or more objects is below the first threshold.

[0052] According to a second aspect of the invention, the above objective is achieved by a camera for facilitating its installation. The camera includes: an image capture module configured to capture a video stream of a scene, the scene including one or more objects, when the camera is installed with a first field of view; a size estimation component configured to estimate the size of the video stream of at least a portion of the one or more objects; and a providing component configured to provide a first edge-filtered version of the video stream to ensure the privacy of the one or more objects depicted in the video stream, provided that the estimated size of the video stream of at least a portion of the one or more objects is higher than a first threshold and lower than a second threshold. In response to the estimated size of the video stream of at least a portion of the one or more objects being higher than a second threshold indicating a risk of privacy breach for at least one object, the camera is configured to enter an enhanced security and privacy operating mode, wherein an alarm generation component is configured to generate an alarm to prompt the installer to reinstall the camera with a second field of view different from the first field of view, and wherein the providing component is configured to prohibit the provision of the first edge-filtered version of the video stream.

[0053] According to a third aspect of the invention, the above objective is achieved by a non-transient computer-readable medium storing computer code instructions that, when executed by a processing-capable device, are suitable for performing the method of the first aspect.

[0054] The second and third aspects can generally have the same features and advantages as the first aspect. It should be further noted that the invention relates to all possible combinations of the features disclosed herein, unless otherwise expressly stated. Attached Figure Description

[0055] The above and additional objectives, features, and advantages of the present invention will be better understood with reference to the accompanying drawings through the following illustrative and non-limiting detailed description of embodiments of the invention, wherein the same reference numerals will be used for similar elements, wherein:

[0056] Figure 1 Exemplary environments in which various embodiments of the present invention may be implemented are illustrated schematically.

[0057] Figure 2 A camera according to an embodiment is illustrated schematically.

[0058] Figure 3 This is a flowchart of a method for facilitating camera installation according to an embodiment.

[0059] Figure 4 The image processing pipeline according to an embodiment is illustrated schematically.

[0060] Figure 5 The illustration schematically depicts image frames from a captured video stream of four objects.

[0061] Figure 6 The illustration schematically depicts privacy-masking image frames from a captured video stream used to ensure the privacy of four objects.

[0062] Figure 7 The illustration schematically depicts the first edge-filtered version of the edge-filtered image frame of a captured video stream of four objects.

[0063] Figure 8 This illustration schematically shows an example of an edge-filtered image frame from a second edge-filtered version of a captured video stream.

[0064] Figure 9 This illustration schematically shows an example of an edge-filtered image frame from another second version of the edge-filtered video stream.

[0065] Figure 10 This illustration schematically shows an example of an edge-filtered image frame from a captured video stream.

[0066] Figure 11 The illustration schematically shows an example of an edge-filtered image frame 1100 of a second edge-filtered version of a captured video stream, where the number of quantization levels is reduced.

[0067] Figure 12 This illustration schematically shows an example of edge-filtered image frame 1200 of a second edge-filtered version of a captured video stream. Detailed Implementation

[0068] The invention will be described more fully below with reference to the accompanying drawings, in which embodiments of the invention are illustrated.

[0069] Figure 1 A schematic diagram of an exemplary environment 100 in which various embodiments of the present invention may be implemented is shown. For example... Figure 1 As can be seen, a scene 102 with an object 104 (e.g., a person walking towards building 106) is captured by a camera 108 mounted on pole 109. It should be noted that the description of scene 102 is a simplified diagram for illustrative purposes only. Scene 102 can be described in a more general sense as any three-dimensional physical space whose size and shape are defined by the field of view of the camera recording the scene. Furthermore, it should be understood that the camera 108 can be mounted in other suitable locations (such as on a building, in a ceiling, on a wall, or on a fence), and therefore, mounting on pole 109 is given only as an example.

[0070] Camera 108 may be a surveillance camera, sometimes referred to as a monitoring camera. Further, the camera may be a fixed camera (e.g., a still camera) or a movable camera (e.g., a pan, tilt, and zoom (PTZ) camera or a wearable camera (BWC)). Camera 108 may be a visible light camera, an infrared (IR) sensitive camera, or a thermal (long-wave infrared (LWIR)) camera.

[0071] like Figure 1 As illustrated, camera 108 is configured to wirelessly transmit a video stream to wired network 112 via radio link 110, and ultimately to a client 114 and / or server 116 that can connect to network 112. It should be understood that many combinations of wireless and wired transmission models are possible, and Figure 1 Only one example is shown.

[0072] Client 114 may have a display where an operator views images and / or video streams from the camera. Typically, client 114 is also connected to server 116 where images and / or video streams can be stored and / or further processed. Client 114 is also typically used to control camera 108, for example, by issuing control commands from the operator at client 114. For example, the operator may instruct the camera to zoom out to a specific detail of scene 102. This could be a situation where, when capturing a video stream using a first field of view (e.g., a first zoom setting), an alarm is generated for the installer when the size of the video stream of a portion of one or more objects 104 is estimated to be above a second threshold, or when capturing a video stream using a second field of view (e.g., a second zoom setting) different from the first field of view, the installer may instruct the camera to zoom out to a portion of one or more objects to reduce the size of the video stream of that portion to below the second threshold. This ensures the privacy of one or more objects. However, there are also situations where the operator cannot control the camera, and camera 108 is fixed and only provides a video stream for the operator to view on client 120. In such a scenario, the installer may need to reinstall camera 108 at a different physical location to obtain a second field of view, in which the size of the video stream is below a second threshold when video is captured using that second field of view. Therefore, reinstalling camera 108 at a different physical location can also achieve the desired security and privacy for one or more objects. The different physical location can be a location at a greater distance from one or more objects compared to the distance when installed with a first field of view.

[0073] Figure 2 An embodiment of camera 108 is illustrated schematically.

[0074] Camera 108 includes a lens 202 that captures scene 102 and projects it onto image sensor 204. Lens 202 and image sensor 204 together form image capture module 206. Image sensor 204 captures image frames that form a video stream. The camera further includes an image processing pipeline (IPP) 208, an encoder 210, and an input / output interface 212.

[0075] Image processing pipeline 208 is configured to perform a series of different operations on image frames received from image sensor 204. These operations may include filtering, demosaicing, color correction, noise filtering for eliminating spatial and / or temporal noise, distortion correction for eliminating effects such as barrel distortion, global and / or local tone mapping, such as enabling imaging of scenes with broad intensities, transformation (e.g., rotation), planar field correction (e.g., for removing vignetting), and overlay applications (e.g., privacy masking, descriptive text, etc.). However, it should be noted that some of these operations, such as barrel distortion correction, rotation, etc., can be performed by one or more modules, components, or circuits located outside image processing pipeline 208, such as in one or more units between image processing pipeline 208 and encoder 210.

[0076] Following image processing pipeline 208, image frames are forwarded to encoder 210, where they are encoded according to an encoding protocol and then forwarded via network 112 to a receiver (e.g., client 114 and / or server 116) using input / output interface 212. It should be noted that... Figure 2 The camera 108 illustrated also includes several other components (such as a processor, memory, etc.) common in conventional camera systems, and their purpose and operation are known to those skilled in the art. For clarity, one can refer to... Figure 2 These components are omitted from the description and specifications.

[0077] Camera 108 may also include data storage 214 for storing data related to video stream capture and data related to methods for facilitating camera 108 mounting. Thus, the data storage may store the captured video stream, a first version and a second version of the video stream, a first threshold and a second threshold, a first field of view and a second field of view, an estimated size of the video stream for at least a portion of one or more objects, to name just a few examples. The data storage may be non-volatile memory, such as an SD card.

[0078] Many traditional video coding formats exist. Some common video coding formats applicable to the various embodiments of the present invention include: JPEG, Motion JPEG (MJPEG), High Efficiency Video Coding (HEVC) (also known as H.265 and MPEG-H Part 2), Advanced Video Coding (AVC) (also known as H.264 and MPEG-4 Part 10), Universal Video Coding (VVC) (also known as H.266, MPEG-1 Part 3 and Future Video Coding (FCV)), VP9, ​​VP10, and AOMedia Video 1 (AV1), to name just a few.

[0079] As described, the input and output interface 212 is configured to serve as a communication interface between the camera 108, the network 112, and one or more receiving devices, such as the client 114 and the server 116. Therefore, the camera 108 can receive instructions from the client 114 via the input and output interface 212 and can transmit video streams to the client and / or the server 116.

[0080] The image processing pipeline 208, encoder 210, and input / output interface 212 can form an image processing and encoding module 214, which is connected to the image capture module 206. The image capture module 206 and the image processing and encoding module 214 can be configured as two independent units positioned at a distance from each other and communicating with each other. In this case, the camera 108 can also be referred to as a camera system. Alternatively, the image capture module 206 and the image processing and encoding module 214 can be configured as a single unit included in the camera 108. Furthermore, the image capture module 206 can be movable (e.g., in the pan and / or tilt direction), while the image processing and encoding module 214 can be fixed.

[0081] Now, refer to Figure 3 The flowchart and reference Figure 2 and Figure 4 Describes a method for facilitating the installation of camera 108. Figure 4 An image processing pipeline 208 according to an embodiment is schematically illustrated. As illustrated, an embodiment of the image processing pipeline 208 includes a size estimation unit 402, a providing unit 404, and an alarm generation unit 406. However, it should be understood that one or more of the size estimation unit 402, the providing unit 404, and the alarm generation unit 406 may be arranged in the camera 108 in another manner. For example, the camera 108 may include one or more of the size estimation unit 402, the providing unit 404, and the alarm generation unit 406 as separate units communicatively connected to each other and communicatively connected to other parts of the camera 108 (such as the image processing pipeline 208, encoder 210, input and output interfaces 212, and data storage 214).

[0082] Furthermore, it should be understood that Figure 3 The method illustrated herein can be performed by different parts of camera 108, some of its steps can be optional, and the steps can be performed in another suitable order. Furthermore, in this disclosure, one or more steps are described as being performed for image frames of the captured video stream. This can be understood as one or more steps being performed on a temporal portion of the video stream.

[0083] In step S302, when camera 108 is mounted with a first field of view (FOV), a video stream of scene 102 is captured. This step is performed by image capture module 206 (e.g., via lens 202 and image sensor 204).

[0084] As previously stated, scenario 102 includes one or more objects 104.

[0085] As previously stated, the term field of view refers to the portion of the scene visible through camera 108 at a specific location and orientation, and with a specific zoom setting of camera 108 within the space of camera 108. The specific location is given by the mounting position, and the orientation is given by the camera's pan and / or tilt settings. Therefore, it should be understood that the first field of view can depend on one or more different camera parameters. For example, the first field of view can depend on the mounting position of camera 108 (such as its height above the ground), the zoom setting of camera 108, the pan setting of camera 108, and / or the tilt setting of camera 108.

[0086] Furthermore, and as previously stated, one or more objects 104 can be any kind of object that could disclose privacy information about the object itself, about the person who owns the object, or about the person associated with the object. Thus, an object can be a person or a part thereof (such as a face or body part), a group of people, a vehicle or a part thereof (such as a vehicle license plate), to name a few examples.

[0087] Figure 5 The image frame 500 of the captured video stream is schematically illustrated, which depicts four objects 104 (e.g., four people 104a, 104b, 104c, 104d) walking on the walkway.

[0088] In step S304, the size of the video stream of at least one portion of one or more objects 104 is estimated. This step can be performed by size estimation component 402. The size of the video stream of the at least one portion of one or more objects can be estimated as the number of pixels of the at least one portion formed along at least one direction of the video stream. For example, the estimated size of the at least one portion can be given as the number of pixels in the vertical (height) direction and / or horizontal (width) direction of an image frame in the video stream. Thus, the estimated size can be given as the length, height, width, or distance between two points as the number of pixels. For example, the estimated size can be the number of pixels between an individual's eyes, or the number of pixels between the left and right sides of an individual's face. However, the estimated size can also be the number of pixels included in the at least one portion (e.g., a region) in an image frame. In an embodiment, where the estimated size is estimated as the size of the video stream of an individual's face, estimation component 402 can include or be connected to a face detector (402a) configured to detect faces in the video stream. However, the estimation component 402 may additionally or alternatively include or be connected to one or more other object detectors (402b) configured to detect other objects or portions thereof (such as vehicles and license plates). Figure 5 In the diagram, bounding boxes 104a', 104b', 104c', and 104d' schematically illustrate the faces of individuals 104a to 104d detected by, for example, face detector 402a. The size of vehicles in the captured video can be detected, for example, by object detector 402b, which estimates the distance between the headlights of vehicles, or by estimating the size of the license plates of vehicles in the captured video stream.

[0089] In step S306, and provided that the estimated size of the video stream of at least one portion of one or more objects 104 is above a first threshold and below a second threshold, a first edge-filtered version of the video stream is provided to ensure the privacy of one or more objects 104 depicted in the video stream. This step is performed by providing component 404.

[0090] In some embodiments, providing a first edge-filtered version of the video stream includes generating the first edge-filtered version of the video stream by applying an edge filter to the video stream to enhance edge points in the video stream and suppress non-edge points in the video stream. The pixel values ​​of the video stream and the pixel values ​​of the first edge-filtered version of the video stream are represented using a first number of quantization levels.

[0091] Providing a first edge-filtered version of the video stream may include transmitting the first edge-filtered version of the video stream to a receiver. As previously described, the receiver may be any device operatively connected to or included within the camera 108 via a wired or wireless connection. Therefore, it should be understood that the first edge-filtered version of the video stream may be transmitted to a client 114, a server 116, and / or a data storage device 214.

[0092] Figure 7 The illustration schematically depicts an edge-filtered image frame 700 of a first edge-filtered version of the captured video stream, depicting four objects 104 (such as four people 104a, 104b, 104c, 104d). The edge-filtered image frame 700 is obtained by applying a Sobel filter to image frame 500 and has the highest possible spatial resolution. The highest possible spatial resolution means that the spatial resolution of the edge-filtered image frame is the same as the corresponding image frame of the captured video stream. In other words, the spatial resolution of the edge-filtered image frame 700 is the same as the spatial resolution of image frame 500. Figure 7 As can be seen, when individuals get closer to the camera, and therefore their faces are enlarged in image frame 700, their privacy is compromised. Therefore, prior to a privacy breach, for example, when the privacy of one or more objects is already at risk of being compromised, the provision of a first edge-filtered version of the captured video stream should be prohibited.

[0093] Therefore, in step S308 and in response to the estimated size of the video stream of at least one portion of one or more objects 104 being higher than a second threshold indicating a risk of privacy breach for at least one object, camera 108 enters an enhanced security and privacy operation mode.

[0094] When operating in enhanced security and privacy mode, in step S310, an alarm is generated to prompt the installer to reinstall camera 108 with a second field of view different from the first field of view, and in step S312, the provision of a first edge-filtered version of the video stream is disabled. The provision of the first edge-filtered version of the video stream can be disabled for a predetermined period of time (e.g., a period of time corresponding to one or more days or the expected period of time required for reinstalling camera 108).

[0095] It should be understood that reinstalling camera 108 to obtain a different second field of view may include changing the camera settings from a first pan level to a different second pan level, and / or from a first tilt setting to a different second tilt setting, and / or from a first zoom setting to a different second zoom setting. The second zoom setting is preferably a reduced setting compared to the first zoom setting. Alternatively or additionally, reinstallation may include moving camera 108 from a first physical location to a different second physical location, i.e., from a pole mounted on the ground to a structure mounted on a tall building.

[0096] After the camera 108 is reinstalled with a second field of view, the method steps can now be repeated from step S302 with a second field of view as the new first field of view (i.e., the first field of view mentioned in steps S302 to S312).

[0097] Step S310 can be performed by alarm generation component 406, and step S312 can be performed by providing component 404.

[0098] In some embodiments, preventing the provision of a first edge-filtered version of the video stream includes preventing the transmission of the first edge-filtered version of the video stream to the receiver. This means that in some embodiments, camera 108 stops recording the first edge-filtered version of the video stream. For example, the first edge-filtered version of the video stream is not stored in data storage 214.

[0099] Alternatively, in some embodiments, providing a first edge-filtered version of the video stream disables the camera from capturing the video stream. This can be when a second version of the video stream is not needed or desired, and therefore, when the first edge-filtered version of the video stream fails to guarantee the privacy of one or more objects, the camera 108 is deactivated. In this case, the image capture module 206 (e.g., lens 202 and / or image sensor 204) can disable the capture of the video stream.

[0100] In some embodiments, providing a first edge-filtered version of blocking the video stream includes blocking the camera's image sensor from transmitting image information used to generate the captured video stream. Therefore, in some cases, the camera 108 is blocked by blocking the image sensor 204 from transmitting image frames used to generate the captured video stream. For example, the image sensor 204 may block the image processing and encoding module 214 from transmitting image frames to the camera 108.

[0101] In optional step S314, the enhanced security and privacy operating mode further includes providing a second version of the video stream that further ensures the privacy of one or more objects 104. This step may be performed during a predetermined time period mentioned in step S312 above, during which the provision of the first edge-filtered version of the video stream is prohibited to await the reinstallation of camera 108. As will be described below, the provision of the second version of the video stream may be implemented in different ways.

[0102] In a first alternative approach, providing a second version of the video stream includes generating the second version of the video stream by applying a privacy mask to at least a portion of the video stream to ensure the privacy of one or more objects 104 depicted in the video stream. The privacy mask can be a color mask or a mosaic (pixel) mask. For example, a privacy mask identifies the live objects and the recorded video stream by comparing the live camera view with a set background scene and applying a dynamic mask to areas of change (especially moving objects). Furthermore, a desired level of masking can be set. A color mask provides maximum privacy protection while allowing viewing of movement. A mosaic mask shows moving objects at a low resolution and, compared to a color mask that applies a single mask color to the entire object, allows for easier differentiation of forms by seeing different colors of the object.

[0103] Figure 6 The illustration schematically depicts a privacy-masking image frame 600 of the captured video stream, which ensures the privacy of four objects 104 (e.g., four people 104a, 104b, 104c, 104d).

[0104] In a second alternative approach, providing a second version of the video stream includes generating a second edge-filtered version of the video stream by reducing the spatial resolution of the video stream before applying an edge filter to reduce the size of the video stream of at least a portion of one or more objects to below a second threshold. As previously described, by reducing the spatial resolution of the video stream before applying the edge filter, such that the size of the video stream of at least a portion of one or more objects is below the second threshold, the privacy of one or more objects is ensured in the second edge-filtered version of the video stream.

[0105] Figure 8An example of an edge-filtered image frame 800 is schematically illustrated as a second edge-filtered version of the captured video stream. The spatial resolution of edge-filtered image frame 800 is half that of edge-filtered image frame 700. As can be seen by comparing edge-filtered image frames 700 and 800, the size of edge-filtered image frame 800 is reduced compared to the image size of edge-filtered image frame 700. However, as can be seen in edge-filtered image frame 800, the privacy of the four individuals is not guaranteed. To ensure privacy, the spatial resolution should be further reduced before applying the edge filter to reduce the size of the video stream of at least one portion of one or more objects to below a second threshold. Figure 9 This schematically illustrates an example of an edge-filtered image frame 900, representing another second version of edge filtering in the captured video stream. The spatial resolution of edge-filtered image frame 900 is one-quarter of the spatial resolution of edge-filtered image frame 700. Figure 9 As illustrated in the figure, the privacy of four people is ensured in edge-filtered image frame 900.

[0106] In a third alternative approach, providing a second version of the video stream includes generating a second version of the video stream as a second edge-filtered version of the video stream by reducing the number of quantization levels used to represent the pixel values ​​of the video stream from a first number of quantization levels to a second number of quantization levels before applying the edge filter to the video stream.

[0107] It should be understood that the pixel values ​​of the second edge-filtered version of the video stream can be represented using the second quantization level.

[0108] However, it should also be understood that the possible number of quantization levels can be increased again to correspond to the first number of quantization levels, and in this case, only the number of quantization levels corresponding to the second number of quantization levels is used to represent the pixel values ​​of the second edge-filtered version of the video stream. For example, if the first number of quantization levels of the video stream corresponds to 256 (8 bits) and the second number of quantization levels is 16 (4 bits), then the second edge-filtered version of the video stream can be a 4-bit video stream, in which 16 quantization levels are used to represent the pixel values, or its number of quantization levels can be increased to 256 (8 bits), but only 16 of the 256 quantization levels are used to represent the pixel values.

[0109] As previously described, by reducing the number of quantization levels used in the video stream from a first number of quantization levels to a lower second number of quantization levels, the graininess in the second edge-filtered version of the video stream increases because the number of intensities to be represented by each quantization level is larger in the second number of quantization levels compared to the number of intensities to be represented by each quantization level in the first number of quantization levels. This greater graininess ensures the privacy of one or more objects depicted in the second edge-filtered version of the video stream.

[0110] Figure 11 An example of an edge-filtered image frame 1100, representing a second edge-filtered version of a captured video stream, is illustrated schematically, wherein the number of quantization levels has been reduced to ensure privacy. In the illustrated figure, the number of quantization levels in edge-filtered image frame 1100 is 8 (3 bits), compared to 256 quantization levels (8 bits) in edge-filtered image frame 700.

[0111] In a fourth alternative approach, providing a second version of the video stream includes: generating a second version of the video stream as a second edge-filtered version of the video stream by reducing the spatial resolution of a first edge-filtered version of the video stream, wherein the size of at least one portion of one or more objects is below a second threshold. As previously described, reducing the spatial resolution of the first edge-filtered version of the video stream reduces the number of pixels in one or more image frames constituting the first edge-filtered version of the video stream. Furthermore, by reducing the spatial resolution of the first edge-filtered version of the video stream such that the size of the second edge-filtered version of the video stream for at least one portion of one or more objects is below the second threshold, privacy of one or more objects is ensured in the second edge-filtered version of the video stream when privacy cannot be guaranteed in the first edge-filtered version of the video stream.

[0112] To ensure the privacy of objects depicted in different regions of the captured video stream, the first image resolution of the first region of the edge-filtered image frame depicting the first object can be different from the second image resolution of the second region of the edge-filtered image frame depicting the second object. The first image resolution of the first region relates to the amount of image detail in the first region, and correspondingly, the second image resolution of the second region relates to the amount of image detail in the second region. The first image resolution and the second image resolution can be obtained by differentiating the first and second regions and then enlarging them to include their original number of pixels. As previously described, reduction decreases the number of pixels, causing a loss of image detail, while enlargement increases the number of pixels but cannot recover the lost image detail. Therefore, by differentiating the first and second regions, different amounts of image detail will be lost in the first and second regions, resulting in different degrees of anonymization. By enlarging the first and second regions to their original number of pixels after reduction, the first and second regions will have their original size in the edge-filtered image frame. Figure 10An example of an edge-filtered image frame 1000 of a second edge-filtered version of a captured video stream is schematically illustrated. In this example, the first resolution of the first region 1002 of the edge-filtered image frame 1000 is half the image resolution of the edge-filtered image frame 700, and the second resolution of the second region 1004 of the edge-filtered image frame 1000 is one-quarter the image resolution of the edge-filtered image frame 700.

[0113] In a fifth alternative, the provision of a second version of the video stream includes: generating a second version of the video stream as a second edge-filtered version of the video stream by representing edge points with edge intensities below an edge intensity threshold in a first edge-filtered version of the video stream as non-edge points in a second edge-filtered version of the video stream. Figure 12 An example of an edge-filtered image frame 1200 of a second edge-filtered version of a captured video stream is schematically illustrated, wherein edge points with edge intensities below an edge intensity threshold are suppressed as non-edge points in the second edge-filtered version of the video stream. This improves the privacy of one or more objects 104. It should be understood that by further increasing the edge intensity threshold, further anonymization of one or more objects 104 will be achieved, and therefore, higher security and privacy for one or more objects 104 will also be obtained.

[0114] In optional step S316, and when the estimated size of at least a portion of the video stream is below a first threshold, the captured video stream is provided. The captured video stream can be provided by transmitting it to a receiver. As previously described, the receiver can be any device operatively connected to or included within camera 108 via a wired or wireless connection. Therefore, it should be understood that the captured video stream can be transmitted to client 114, server 116, and / or data storage 214.

[0115] The embodiments also relate to a camera 108 for facilitating its installation. The camera 108 includes an image capture module 206 configured to capture a video stream of a scene 102 when the camera 108 is mounted with a first field of view, wherein the scene 102 includes one or more objects 104.

[0116] The camera 108 also includes a size estimation component 402, configured to estimate the size of a video stream of at least one portion of one or more objects 104.

[0117] Furthermore, the camera 108 includes a providing component 404 configured to provide a first edge-filtered version of the video stream to ensure the security and privacy of one or more objects depicted in the video stream, provided that the estimated size of the video stream of at least one portion of one or more objects 104 is above a first threshold and below a second threshold.

[0118] The providing component 404 is configured to detect and enhance edge points and suppress non-edge points in the captured video stream, and is configured to provide an edge-filtered version of the captured video stream, such as a first edge-filtered version and a second edge-filtered version of the captured video stream as described above. The providing component 404 may be or may include an edge detector configured to provide an edge-filtered version of the captured video stream.

[0119] In addition, component 404 is configured to disable the provision of a first edge-filtered version of the video stream when camera 108 is in enhanced security and privacy operating mode.

[0120] In some embodiments, the providing component 404 is configured to provide a second version of the captured video stream, wherein the second version may be a privacy-masking version or a second edge-filtered version of the captured video stream.

[0121] The providing component 404 can also be configured to provide the captured video stream when the estimated size of the video stream of at least one portion of one or more objects 104 is below a first threshold.

[0122] Camera 108 is configured to enter an enhanced security and privacy operation mode in response to an estimated size of a video stream of at least one portion of one or more objects 104 exceeding a second threshold indicating a risk of privacy breach for at least one object.

[0123] Furthermore, the camera 108 includes an alarm generation component 406, configured to generate an alarm when the camera 108 is in an enhanced security and privacy operating mode, to prompt the installer to reinstall the camera in a second field of view different from the first field of view.

[0124] The embodiments also relate to a non-transient computer-readable medium having computer code instructions stored thereon, which, when executed by a processing-capable device, are suitable for performing the embodiments described herein.

[0125] As described above, camera 108 can be configured to implement methods for facilitating its installation. For this purpose, camera 108 may include circuitry configured to implement the various method steps described herein.

[0126] In a hardware implementation, the circuitry can be dedicated and specifically designed to implement one or more method steps. The circuitry can take the form of one or more integrated circuits, such as one or more application-specific integrated circuits (ASICs) or one or more field-programmable gate arrays (FPGAs). By way of example, camera 108 can therefore include circuitry for estimating the size of a video stream of at least a portion of one or more objects during use, and for providing a first edge-filtered version of the video stream to ensure the privacy of one or more objects depicted in the video stream when the estimated size of the video stream of at least a portion of one or more objects is above a first threshold and below a second threshold during use. Camera 108 can further include circuitry for entering an enhanced security and privacy operating mode when the estimated size of the video stream of at least a portion of one or more objects is above a second threshold indicating a risk of privacy breach for at least one object during use. Furthermore, camera 108 can include circuitry for generating an alarm during use and when in enhanced security and privacy mode to prompt the installer to reinstall the camera in a second field of view different from the first field of view and to disable the provision of the first edge-filtered version of the video stream.

[0127] In a software implementation, the circuitry may alternatively take the form of a processor (such as a microprocessor), associated with computer code instructions stored on a (non-transient) computer-readable medium (such as non-volatile memory), causing the camera 108 to perform any of the methods disclosed herein. Examples of non-volatile memory include read-only memory, flash memory, ferroelectric RAM, magnetic computer storage devices, and optical discs. In the software implementation, each step of the above method steps can therefore correspond to a portion of computer code instructions stored on a computer-readable medium, which, when executed by the processor, cause the camera 108 to perform any of the methods disclosed herein.

[0128] It should be understood that there can also be a combination of hardware and software implementations, meaning that some steps are implemented in hardware while others are implemented in software.

[0129] It should be understood that those skilled in the art can modify the above embodiments in many ways and still utilize the advantages of the invention as shown in the above embodiments. For example, the camera 108 need not be a single unit comprising the image capture module 206 and the image processing and encoding module 214 in one location, but it can be a virtual unit, wherein the image capture part 206 operates together with the image processing and encoding module 214 but they are provided in different locations. Furthermore, the size estimation component 402, the providing component 404, and the alarm generation component 406 need not be located in the image processing pipeline 208, but can be set as independent units of the image processing and encoding module 214 and can be configured to communicate with the image processing pipeline 208, the encoder 210, the input and output interface 212, and the data storage 214. Therefore, the invention should not be limited to the embodiments shown, but should be defined only by the appended claims. Additionally, as those skilled in the art should understand, the illustrated embodiments can be combined.

Claims

1. A method performed in a camera for facilitating the installation of the camera, the method comprising: When the camera is mounted with the first field of view, a video stream of a scene (S302) is captured, the scene including one or more people. Detect faces in the video stream; Estimating (S304) the size of each detected face in the video stream, including: The estimated size of each detected face in the video stream is determined to be higher than a first threshold and lower than a second threshold, wherein the first threshold is a higher threshold that allows the video stream to be provided as is without disclosing the privacy of the one or more persons depicted in the video stream, and wherein the second threshold is a higher threshold that allows a first edge-filtered version of the video stream to be provided without disclosing the privacy of the one or more persons depicted in the first edge-filtered version of the video stream. Provide (S306) the first edge-filtered version of the video stream as long as the estimated size of each detected face in the video stream is higher than the first threshold and lower than the second threshold, to ensure the security and privacy of the one or more persons depicted in the video stream; The estimated size of the detected face in the video stream is higher than the second threshold; and In response to a detected face whose estimated size in the video stream exceeds a second threshold indicating a risk of privacy breach in a first edge-filtered version of the video stream: An alarm is generated (S310) to prompt the installer to reinstall the camera in a second field of view, different from the first field of view, in which the size of the detected face in the video stream will become below the second threshold, and Temporarily disable (S312) the provision of the first edge-filtered version of the video stream for the duration required for the camera to be reinstalled with the second field of view.

2. The method according to claim 1, further comprising: During the time period required to reinstall the camera in the second field of view, a second version of the video stream is provided (S314) to further ensure the privacy of the one or more persons compared to the first edge-filtered version of the video stream.

3. The method according to claim 2, wherein, The second version of providing (S314) the video stream includes: The second version of the video stream is produced by applying a privacy shield to at least a portion of the video stream to ensure the privacy of the one or more persons depicted in the video stream.

4. The method according to any one of claims 1 to 3, wherein, The first edge-filtered version of the video stream provided (S306) includes: The first edge-filtered version of the video stream is generated by applying an edge filter to the video stream to enhance edge points in the video stream and suppress non-edge points in the video stream, wherein the pixel values ​​of the video stream are represented using a first quantization level.

5. The method according to claim 4, wherein, The second version of providing (S314) the video stream includes: The second version of the video stream is generated as the second edge-filtered version of the video stream in the following manner: Before applying the edge filter to the video stream, the spatial resolution of the video stream is reduced to reduce the size of the detected faces of the one or more persons in the video stream to below the second threshold.

6. The method according to claim 4, wherein, The second version of providing (S314) the video stream includes: The second version of the video stream is generated as the second edge-filtered version of the video stream in the following manner: Before applying the edge filter to the video stream, the number of quantization levels used to represent the pixel values ​​of the video stream is reduced from the first number of quantization levels to the second number of quantization levels.

7. The method according to any one of claims 1 to 3, wherein, The second version of providing (S314) the video stream includes: The second version of the video stream is generated as the second edge-filtered version of the video stream in the following manner: The spatial resolution of the first edge-filtered version of the video stream is reduced to obtain a second edge-filtered version of the video stream, wherein the size of the detected faces of the one or more persons in the video stream is below a second threshold.

8. The method according to any one of claims 1 to 3, wherein, The second version of providing (S314) the video stream includes: The second version of the video stream is generated as the second edge-filtered version of the video stream in the following manner: Edge points with edge strength below the edge strength threshold in the first edge-filtered version of the video stream are represented as non-edge points in the second edge-filtered version of the video stream.

9. The method according to any one of claims 1 to 3, wherein, The first edge-filtered version of the video stream provided (S306) includes: The first edge-filtered version of the video stream is transmitted to the receiver; and The provision of the first edge-filtered version of the video stream that is prohibited includes: Transmission of the first edge-filtered version of the video stream to the receiver is prohibited.

10. The method according to claim 1, wherein, Prohibiting (S312) the provision of the first edge-filtered version of the video stream includes: The camera (108) is prohibited from capturing the video stream.

11. The method according to claim 1, wherein, Prohibiting (S312) the provision of the first edge-filtered version of the video stream includes: The image sensor (204) of the camera (108) is prohibited from transmitting image frames used to generate the captured video stream.

12. The method according to any one of claims 1 to 3, further comprising: When the estimated size of each detected face of the one or more persons in the video stream is lower than the first threshold, the captured video stream is provided (S316).

13. A camera for facilitating camera installation, the camera comprising: The image capture module (206) is configured to capture a video stream of a scene, including one or more people, when the camera is mounted with a first field of view. A face detector (402a) is configured to detect faces in the video stream; A size estimation component (402) is configured to estimate the size of each detected face in the video stream. The camera is configured to determine that the estimated size of each detected face in the video stream is higher than a first threshold and lower than a second threshold, wherein the first threshold is a higher threshold that allows the video stream to be provided as is without disclosing the privacy of the one or more persons depicted in the video stream, and wherein the second threshold is a higher threshold that allows the video stream to be provided with a first edge-filtered version without disclosing the privacy of the one or more persons depicted in the first edge-filtered version of the video stream; The camera includes: A component (404) is configured to provide a first edge-filtered version of the video stream to ensure the privacy of the one or more persons depicted in the video stream, provided that the estimated size of each determined face in the video stream is higher than a first threshold and lower than a second threshold. The camera is configured to determine that the estimated size of a detected face in the video stream is higher than a second threshold; and The camera includes an alarm generation component (406), and wherein, in response to an estimated size of a detected face in the video stream exceeding a second threshold indicating a risk of privacy breach for at least one person: The alarm generation component (406) is configured to generate an alarm to prompt the installer to reinstall the camera in a second field of view different from the first field of view, in which the size of the detected face in the video stream will become below the second threshold, and The providing component (404) is configured to temporarily disable the provision of the first edge-filtered version of the video stream for the duration required for the camera to be reinstalled with the second field of view.

14. A non-transient computer-readable medium storing computer code instructions that, when executed by a processing-capable device, are adapted to perform the method according to any one of claims 1 to 12.

Images