Sensor device and encryption method

Abstract

The present invention realizes image encryption, in which excessive hiding of information is prevented, while achieving cost reduction by eliminating the need for the image receiving side to take measures to prevent leakage of personal information. The sensor device according to the present invention includes: an array sensor, in which a plurality of pixels each having a light receiving element for visible light or invisible light are arranged in one or two dimensions; a detection unit, which detects an area of ​​a target projected in an image as a target area based on an image signal obtained by imaging with the array sensor; and an encryption unit, which performs encryption for the target area in the image signal based on information of the target area detected by the detection unit. This makes it possible to encrypt the image signal obtained by imaging with the array sensor in the image sensor, at least to the extent that an individual cannot be identified. In addition, depending on the state of saving the decryption key on the receiving side of the image, a portion of the content of the image can be visually identified while hiding personal information.

Description

Technical Field

[0001] The present technology relates to a sensor device and an encryption method, and in particular to a technical field related to encryption of image signals obtained by an array sensor. Background Art

[0002] For example, based on image signals obtained by imaging with a surveillance camera, it is possible to use an image analysis device external to the camera (e.g., a cloud server, etc.) to analyze the attributes and behavior of objects (e.g., people) projected in the image. For example, it is conceivable that by analyzing images captured by a surveillance camera installed in a store, attribute information such as the gender and age of customers can be obtained, and the behavior (flow) of customers in the store can be identified and used for marketing purposes.

[0003] Furthermore, as a related prior art, the following Patent Document 1 discloses encryption of a captured image.

[0004] Prior art literature

[0005] Patent Literature

[0006] Patent Document 1 JP 2009-027333 A Summary of the Invention

[0007] [Technical Issues]

[0008] Here, in a case where a person himself or information capable of identifying the person is projected onto an image captured by a camera, in the context of strong demands for personal information protection in recent years, the party that manages the image captured by the camera (i.e., the party that receives the camera image) needs to ensure a high level of security to prevent leakage of personal information, which leads to increased costs.

[0009] The present technology has been proposed in view of the above circumstances, and its object is to realize image encryption in which excessive concealment of information is prevented while achieving cost reduction by eliminating the need for the image receiving side to take measures to prevent leakage of personal information.

[0010] [Solution to the problem]

[0011] A sensor device according to the present technology includes: an array sensor in which a plurality of pixels, each having a light receiving element for visible light or invisible light, are arranged one-dimensionally or two-dimensionally; a detection unit that detects, based on an image signal obtained by imaging with the array sensor, an area of ​​a target projected in an image as a target area; and an encryption unit that performs encryption for the target area in the image signal based on information of the target area detected by the detection unit.

[0012] Therefore, the image signal obtained by imaging with the array sensor can be encrypted in the image sensor, at least to the extent that no individual can be identified.

[0013] Furthermore, depending on the state in which the decryption key is saved on the receiving side of the image, a portion of the content of the image can be visually recognized while concealing personal information.

[0014] In the sensor device according to the present technology described above, it is conceivable that, in encryption of the image signal, the encryption unit encrypts the entire image based on the first encryption key and encrypts the target area based on the first encryption key and a second encryption key different from the first encryption key.

[0015] Therefore, the portion of the image that can be decrypted is distinguished by the decryption key stored on the receiving side of the image. Specifically, if only the decryption key corresponding to the first encryption key is stored, only the area of ​​the image excluding the target area can be decrypted, while if the decryption keys corresponding to the first encryption key and the second encryption key are stored, the entire image including the target area can be decrypted.

[0016] In the sensor device according to the present technology described above, it is conceivable that the detection unit performs processing of identifying a specific portion of the target, and the encryption unit performs encryption based on different encryption keys for the specific portion of the target area and other areas.

[0017] Therefore, the concealment level of the target can be changed according to the decryption key stored on the image receiving side. For example, if the target is a person, the concealment level can be classified into, for example, concealment of the whole body and concealment of only the face.

[0018] In the sensor device according to the present technology described above, it is conceivable to include: an analysis unit that analyzes an attribute or action of a target; and an output unit that outputs information indicating a result of the analysis performed by the analysis unit.

[0019] Therefore, even in the case where the receiving side of the image does not have the decryption key, analysis results of the attributes and actions of the target can be provided.

[0020] In the sensor device according to the present technology described above, it is conceivable that the detection unit performs detection of the target by decrypting the image signal encrypted based on the first encryption key using the just-in-time method.

[0021] Therefore, the possibility of leakage of the image signal in the plain text state can be reduced.

[0022] In the sensor device according to the present technology described above, it is conceivable that the encryption unit encrypts the image signal based on a synthetic key obtained by synthesizing a plurality of encryption keys.

[0023] Therefore, the number of execution times of encryption processing required to achieve encryption in which the concealment level of personal information gradually changes can be reduced.

[0024] In the sensor device according to the present technology described above, it is conceivable that the encryption unit performs encryption for the target area while tracking the target.

[0025] Therefore, in the case where the image to be encrypted is a moving image, the moving object can be appropriately concealed.

[0026] In the sensor device according to the present technology described above, it is conceivable that the encryption unit encrypts the image signal using a stream cipher.

[0027] Therefore, no pre-processing of image signal encryption is required.

[0028] In the sensor device according to the present technology described above, it is conceivable that the array sensor, the detection unit, and the encryption unit are configured in one package.

[0029] Therefore, tamper resistance can be improved in hardware.

[0030] In the sensor device according to the present technology described above, it is conceivable that the encryption unit generates an encryption key based on a photoelectric random number obtained based on the photoelectric conversion performed by the array sensor, and performs encryption of the image signal based on the generated encryption key.

[0031] Therefore, encryption of the encryption key that is more difficult to decrypt than the case of using pseudo-random numbers can be achieved.

[0032] In the sensor device according to the present technology described above, it is conceivable that the encryption unit encrypts the read signals from the pixels of the array sensor.

[0033] Therefore, it is possible to prevent the plaintext image signal from being stored in the memory while being encrypted.

[0034] The encryption method according to the present technology is an encryption method including: detecting an area of ​​a target projected in an image based on an image signal obtained by imaging with an array sensor, as a target area, in which a plurality of pixels each having a light receiving element for visible light or invisible light are arranged one-dimensionally or two-dimensionally in the array sensor; and performing encryption for the target area in the image signal based on information of the detected target area.

[0035] With this encryption method, the same operation as that of the sensor device according to the present technology can also be obtained. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 is a block diagram of a sensor device according to an embodiment of the present technology;

[0037] Figure 2 is a diagram illustrating an example of a method of generating an encryption filter (encryption key) based on a photoelectric random number;

[0038] Figure 3 is a diagram showing an encrypted image of a read signal performed by an amplitude control circuit according to an embodiment;

[0039] Figure 4 is a diagram showing a configuration example of a sensor device in the case of encrypting a read signal generated as a digital signal;

[0040] Figure 5 is a diagram showing a structural example of a sensor device of an embodiment;

[0041] Figure 6 is a diagram showing another example of the structure of the sensor device of the embodiment;

[0042] Figure 7 is a diagram showing still another example of the structure of the sensor device of the embodiment;

[0043] Figure 8 is a flowchart showing the procedure of processing to be performed in order to implement encryption of the first embodiment;

[0044] Figure 9 is an explanatory diagram of output of analysis result data in the first embodiment;

[0045] Figure 10 is a diagram for explaining an example of a method of securely transmitting a key for decryption to a receiving side of an image;

[0046] Figure 11 is a diagram showing an image that is gradually encrypted when the target class is a person;

[0047] Figure 12 is a diagram showing an image that is gradually encrypted when the target class is a vehicle;

[0048] Figure 13 is a diagram for explaining an example of a specific method of step-by-step encryption;

[0049] Figure 14 is an illustration of an example of hidden level changes;

[0050] Figure 15 is a diagram for explaining an example of a specific encryption method of the second embodiment;

[0051] Figure 16 is a diagram for explaining a modified example of ROI in the second embodiment;

[0052] Figure 17is a flowchart showing processing corresponding to processing from imaging of a seed frame to storage of a random number as a source of an encryption key in the second embodiment;

[0053] Figure 18 is a flowchart showing a process of encrypting a target image based on an encryption key generated in the second embodiment;

[0054] Figure 19 is an explanatory diagram of output of analysis result data in the second embodiment;

[0055] Figure 20 is an explanatory diagram of an example of generating an encryption key, in which a random value is assigned to each color cell. DETAILED DESCRIPTION

[0056] Hereinafter, the embodiments will be described in the following order.

[0057] <1. First embodiment>

[0058] [1-1. Configuration of sensor device]

[0059] [1-2. About random numbers used for encryption]

[0060] [1-3. Regarding the signal to be processed]

[0061] [1-4. Regarding tamper resistance]

[0062] [1-5. Processing procedures]

[0063] [1-6. Example of output of analysis information]

[0064] [1-7. Regarding key transmission]

[0065] <2. Second embodiment>

[0066] [2-1. Encryption method of the second embodiment]

[0067] [2-2. Processing Procedure]

[0068] [2-3. Example of output of analysis information]

[0069] <3. Modifications>

[0070] <4. Overview of Embodiments>

[0071] <5. This technology>

[0072] In the embodiments described below, a sensor device 1 is exemplified as an image sensor having an array of light-receiving elements and outputting an image signal as a detection signal. In particular, the sensor device 1 of this embodiment is a device having an object detection function through image analysis and can be referred to as a smart array sensor.

[0073] <1. First embodiment>

[0074] [1-1. Configuration of sensor device]

[0075] Figure 1 FIG. 1 shows a configuration example of the sensor device 1. Figure 1 Also shown are an external processor 11 and an external sensor 12 as external devices that perform data communication with the sensor device 1. It is assumed that the external processor 11 is a general-purpose processor connected to the sensor device 1 for communication.

[0076] As hardware, the sensor device 1 includes components that function as an image sensor device, a storage device such as DRAM (Dynamic Random Access Memory), and an AI (Artificial Intelligence) function processor. These three components are configured as an integrated device by forming a three-layer stacked structure, a so-called flat structure with one layer, or a two-layer stacked structure (for example, the memory device and the AI ​​function processor are on the same layer).

[0077] like Figure 1 As shown, the sensor device 1 has an array sensor 2 , an amplitude control circuit 10 , an ADC (Analog-to-Digital Converter) / pixel selector 3 , a buffer 4 , a logic unit 5 , a memory 6 , an interface unit 7 and a calculation unit 8 .

[0078] The array sensor 2 is configured such that the detection elements are formed of light-receiving elements for visible or invisible light, and a plurality of pixels, each having a light-receiving element, are arranged one-dimensionally or two-dimensionally. For example, a large number of pixels are arranged two-dimensionally in the row and column directions, and a two-dimensional image signal is output due to photoelectric conversion in the light-receiving element of each pixel.

[0079] In addition, in the following description, the array sensor 2 used as an image sensor outputs a two-dimensional image signal, but the array sensor 2 in the sensor device 1 can be configured by a sensor array module provided with an acoustic wave detection element, a sensor array module provided with a tactile information detection element, etc.

[0080] The amplitude control circuit 10 performs amplitude control on the electric signal (analog signal) photoelectrically converted by the array sensor 2. In this example, the amplitude control circuit 10 is configured so that the amplification factor can be changed based on an instruction from the calculation unit 8, which will be described later.

[0081] The electric signal photoelectrically converted by the array sensor 2 is input to the ADC / pixel selector 3 via the amplitude control circuit 10. The ADC / pixel selector 3 converts the input electric signal as an analog signal into digital data and outputs an image signal as digital data.

[0082] Furthermore, the ADC / pixel selector 3 has a pixel selection function for the pixels (imaging elements) of the array sensor 2. Therefore, it is possible to obtain a photoelectric conversion signal and output it as digital data only for selected pixels in the array sensor 2. That is, the ADC / pixel selector 3 generally outputs the photoelectric conversion signal as digital data for all effective pixels constituting one frame of image, but it is also possible to output the photoelectric conversion signal as digital data for only selected pixels.

[0083] The ADC / pixel selector 3 acquires an image signal in units of frames, and the image signal of each frame is temporarily stored in the buffer 4 , read out at an appropriate time, and used for processing in the logic unit 5 .

[0084] The logic unit 5 performs various necessary types of signal processing (image processing) on ​​the input image signal of each frame.

[0085] For example, it is assumed that the logic unit 5 performs image quality adjustment through processing such as color correction, gamma correction, color grayscale processing, gain processing, edge enhancement processing, etc. In addition, it is assumed that the logic unit 5 performs processing for changing data size, such as data compression processing, resolution conversion, frame rate conversion, etc.

[0086] Parameters for each type of processing performed by the logic unit 5 are set. For example, there are setting values ​​such as color and brightness correction coefficients, gain values, compression rates, and frame rates. The logic unit 5 uses the parameters set for each type of processing to perform the necessary processing. In this embodiment, the calculation unit 8 can set these parameters.

[0087] The image signal processed by the logic unit 5 is stored in a memory 6 composed of, for example, a DRAM.

[0088] Furthermore, it is assumed that the memory 6 is a DRAM, SRAM (static random access memory), MRAM (magnetoresistive random access memory), or the like. MRAM is a memory that stores data magnetically, and is known to use a TMR (tunnel magnetoresistive) element instead of a magnetic core. The TMR element has an extremely thin insulating layer composed of a few atoms sandwiched between magnetic materials, and its resistance changes according to the magnetization direction of the magnetic material layer. Even when the power is turned off, the magnetization direction of the TMR element does not change, thereby producing a non-volatile memory. Since a larger write current is required as the size becomes smaller, in order to make the memory cell smaller, an STT-MRAM method using a spin injection magnetization reversal method (spin torque transfer (STT)) is known, in which electrons with uniform spin flow for writing without using a magnetic field. Specific examples of the memory 6 may also include other storage elements.

[0089] The image signal stored in the memory 6 is transmitted and outputted to the external processor 11 or the like by the interface unit 7 at a desired timing.

[0090] The external processor 11 performs image analysis and image recognition processing on the image signal transmitted from the sensor device 1 to perform necessary object detection and the like.

[0091] The external processor 11 may also refer to detection information of the external sensor 12 .

[0092] Here, the external processor 11 is considered to be connected to the sensor device 1 in a wired or wireless manner.

[0093] Furthermore, the imaging device including the sensor device 1 may have a network communication function, and may also adopt a configuration in which the image signal (captured image signal) obtained by the sensor device 1 can be transmitted to a computer device (e.g., a cloud server, etc.) outside the imaging device via a network (e.g., via a network such as the Internet or a local area network (LAN)). In this case, the external processor 11 may be, for example, a processor in a cloud computing system.

[0094] The computing unit 8 is constructed, for example, from an AI processor. Furthermore, as shown in the figure, the computing unit 8 includes a key frame selection unit 81, an object region identification unit 82, a category identification unit 83, a parameter selection unit 84, an encryption control unit 85, and an unauthorized access detection unit 86. Furthermore, these computing functions may be implemented by multiple processors.

[0095] The key frame selection unit 81 performs a process of selecting key frames as moving images from frames of an image signal according to a predetermined algorithm or instruction.

[0096] The object area identification unit 82 performs detection of an area of ​​an object, which serves as a candidate for detection and identification processing of an area (bounding box) of a detection target object in an image (frame) surrounding the detection target object, with respect to a frame of an image signal photoelectrically converted by the array sensor 2 and acquired by the ADC / pixel selector 3.

[0097] The object detected from the image signal is an object that can be used as a detection target for the purpose of recognition from the image. Although the type of object to be detected varies depending on the detection purpose, processing capabilities, application type, etc. of the sensor device 1 and the external processor 11, any object can be a detection target object. Although only a few are shown, any type of object is applicable, for example, animals, mobile objects (cars, bicycles, airplanes, etc.), natural objects (vegetables, plants, etc.), industrial products and parts, buildings, facilities, mountains, oceans, rivers, stars, the sun, clouds, etc.

[0098] Furthermore, the object region identifying unit 82 in this example performs a process of calculating an ROI (Region of Interest) which is region information indicating a region (region of interest) that should be a processing target based on a bounding box.

[0099] The category identification unit 83 performs classification of the object detected by the object region identification unit 82 .

[0100] A category is information representing an object category, and is a category obtained by classifying the object to be detected, for example, "person", "car", "airplane", "ship", "truck", "bird", "cat", "dog", "deer", "frog", "horse", etc.

[0101] The parameter selection unit 84 stores parameters for signal processing according to each category, and selects one or more parameters corresponding thereto using the category, bounding box, etc. of the detection object identified by the category identification unit 83. Furthermore, one or more parameters are set in the logic unit 5.

[0102] The various functions executed by computing unit 8 are not typically performed by an image sensor. In this embodiment, object detection, category recognition, and control based on these functions are performed by the image sensor. Therefore, the image signal supplied to external processor 11 can be used as a suitable signal for detection purposes, and the amount of data can be appropriately reduced without degrading detection performance.

[0103] The encryption control unit 85 performs control so that an image signal obtained by imaging with the array sensor 2 is encrypted. Furthermore, a specific example of the process for encrypting such an image signal performed by the encryption control unit 85 will be described again.

[0104] The unauthorized access detection unit 86 detects unauthorized access from outside the sensor device 1. Specifically, the unauthorized access detection unit 86 in this example detects unauthorized access to data stored in the memory 6 from the outside.

[0105] In the case where the unauthorized access detection unit 86 detects unauthorized access, log information thereof (eg, information indicating the detection date and time, the type of unauthorized access, etc.) is recorded in a predetermined area of ​​the memory 6 .

[0106] [1-2. About random numbers used for encryption]

[0107] Although the sensor device 1 of this embodiment encrypts image signals, conventionally, pseudo-random numbers generated by software have often been used as random numbers for encryption. However, since pseudo-random numbers are generated by algorithms that calculate numerical values ​​and cannot generate true random numbers, there is a risk that the encryption key could be decrypted and copied.

[0108] With this in mind, in this example, optoelectronic random numbers are used to generate encryption keys.

[0109] The photoelectric random number is a random number obtained based on the photoelectric conversion performed by the array sensor 2. Specifically, in this embodiment, the value of the electric signal of each pixel obtained by the photoelectric conversion of the array sensor 2 is obtained as the photoelectric random number to generate the encryption key.

[0110] Figure 2 An example of a method of generating an encryption filter (encryption key) based on optoelectronic random numbers is shown.

[0111] First, the left side of the figure shows the electric signal value of each pixel obtained by photoelectric conversion of the array sensor 2. Specifically, in this example, each pixel value (luminance value) of an image (still image) obtained by imaging by the array sensor 2 is used as a photoelectric random number.

[0112] Hereinafter, a frame image captured in order to obtain a photoelectric random number (in other words, a frame image that is a source of the photoelectric random number) is referred to as a “seed frame”.

[0113] In this example, the value of the electrical signal of each pixel itself is not used as an encryption key. However, as shown on the right side of the figure, an encryption key is generated in a format in which at least some of the values ​​of the electrical signal of each pixel are assigned to pixel positions different from the pixel positions from which the electrical signal values ​​were obtained. In other words, the pixel positions are transformed relative to the values ​​of the electrical signal of each pixel obtained as photoelectric random numbers to generate the encryption key.

[0114] Therefore, compared with the case of using an encryption key, in which the electric signal value of each pixel area is directly assigned to the pixel position where the electric signal value is obtained, it is difficult to decrypt the encryption key and security can be improved.

[0115] Here, when generating the encryption key, a predetermined algorithm can also be used to modulate the value of the electrical signal of each pixel and use it. For example, a value obtained by multiplying the electrical signal value of each pixel by a predetermined coefficient can be used as the random value for the pixel. Alternatively, if the value of the electrical signal of each pixel includes a value after a decimal point, a method can also be used in which a value obtained by converting the value of the last few decimal places to an integer is used as the random value for the pixel position.

[0116] Furthermore, when generating an encryption key, it is not necessary to transform the pixel position as described above, and the value of the electric signal of each pixel itself can be used as the encryption key.

[0117] [1-3. Regarding the signal to be processed]

[0118] Conventionally, in the case of encrypting an image signal obtained by imaging by the array sensor 2 , the image signal read from the array sensor 2 is generally temporarily stored in a memory in a plaintext state, and encryption is performed on the stored image signal.

[0119] However, when this encryption method is adopted, hacker attacks are possible by intentionally causing errors in encryption using malware, etc., outputting the contents of the memory to a dump file, and copying the plaintext placed in the memory.

[0120] Therefore, in the present embodiment, the read signals from the pixels of the array sensor 2 are encrypted so that the plaintext image signals are not stored in the memory.

[0121] Specifically, in this example, Figure 1 The amplitude control circuit 10 shown performs the operation according to the corresponding Figure 2 The amplitude of the coefficients of the encryption key used for reading signals from the pixels of the array sensor 2 is shown to be controlled, thereby achieving encryption of the read signals.

[0122] Figure 3 is a diagram showing an encrypted image of a read signal performed by the amplitude control circuit 10 .

[0123] As shown in the figure, the read signal (in this case, the charge signal) from each pixel in the array sensor 2 is multiplied by a coefficient corresponding to the encryption key through an amplifier provided in the amplitude control circuit 10. Figure 1In the illustrated sensor device 1 , after the amplitude of the read signal of each pixel is controlled in an analog signal state, the signal is A / D-converted by the ADC / pixel selector 3 and stored in the memory 6 via the buffer 4 and the logic unit 5 .

[0124] The encryption control unit 85 sets a coefficient corresponding to the encryption key in the above-mentioned amplifier, and thus the read signal from each pixel in the array sensor 2 is encrypted.

[0125] also, Figure 3 It is only an image diagram, and it is not necessary to provide an amplifier for each pixel in the amplitude control circuit 10. For example, in the case of performing batch reading as in a CCD (Charge Coupled Device) image sensor, the amplifier included in the amplitude control circuit 10 may be one amplifier shared by each pixel. In this case, the amplitude control of each pixel is performed by time division.

[0126] Here, it can be said that the read signal of each pixel is a signal constituting an image signal. In other words, the read signal of each pixel belongs to the image signal in the sense that it constitutes a part of the image signal.

[0127] Although an example of encrypting a read signal which is an analog signal is given above as an example of encrypting the read signal, it is also possible to encrypt a read signal which is a digital signal after A / D conversion.

[0128] Figure 4 A configuration example of the sensor device 1 in the case where a read signal is encrypted into a digital signal is shown.

[0129] In this case, the sensor device 1 is provided with an amplitude control circuit 10A, instead of the amplitude control circuit 10 , which controls the amplitude of the read signal converted into a digital signal by the ADC / pixel selector 3 .

[0130] Furthermore, since the processing performed by the encryption control unit 85 in this case is the same except that the setting destination of the coefficient of each pixel corresponding to the encryption key is changed from the amplitude control circuit 10 to the amplitude control circuit 10A, its repeated description will be omitted.

[0131] Here, when the analog read signal is encrypted as described above, it is very difficult to obtain the analog signal from the outside without authorization, and thus security can be improved.

[0132] Furthermore, in the case where the analog read signal is encrypted, there is a problem that the reproducibility of an image obtained by decrypting the encrypted image may be degraded.

[0133] However, for example, in the case of analyzing attributes and behaviors of objects (eg, people) using an object image, if the object can be detected and analyzed, image reproducibility is considered sufficient and there is no practical problem.

[0134] On the other hand, in the case where the digital read signal is encrypted, the accuracy of the encryption process can be improved, and image reproducibility can be improved.

[0135] Here, the encryption performed on the read signal as described above is an encryption using a stream cipher. The stream cipher is an encryption method that encrypts plain text in predetermined data units (eg, bit units or byte units).

[0136] In stream ciphers, there is no need to make the data length of the encryption target signal consistent, so there is no need to perform pre-encryption processing on the target signal. Therefore, by using stream ciphers, the encryption process can be accelerated.

[0137] [1-4. Regarding tamper resistance]

[0138] In the sensor device 1 of this example, as Figure 5 As shown, each chip of the array sensor 2, memory 6 and computing unit 8 is packaged in one package to improve the tamper resistance in hardware. Figure 5 In the example of , a chip serving as the memory 6 is stacked on a chip serving as the computing unit 8 , and a chip serving as the array sensor 2 is further stacked on a chip serving as the memory 6 .

[0139] In this example, an encryption unit that encrypts a read signal is formed in a chip, for example, as the array sensor 2 .

[0140] Furthermore, an encryption control unit 85 is included in the chip as the calculation unit 8 , and generates an encryption key based on the photoelectric random number and causes the encryption unit to perform encryption based on the encryption key.

[0141] In this example, electrical connection between chips is achieved using Cu-Cu connection, in which Cu (copper) pads are connected, and the electrically connected portions of the chips are destroyed when attempting to disassemble the sensor device 1. That is, this makes it possible to improve tamper resistance in terms of hardware.

[0142] Figure 6 Another example of the structure of the sensor device 1 is shown. Figure 5 The difference is that the vertical relationship between the computing unit 8 and the memory 6 is interchanged.

[0143] Figure 7 Another example of the structure of the sensor device 1 is shown, and Figure 5The difference is that a plurality of chips of the memory 6 are stacked (two layers in the example of this figure).

[0144] In addition, although not shown, the sensor device 1 can have a two-layer structure in which the memory 6 is formed in the same layer as the computing unit 8, or can have a one-layer structure in which the array sensor 2, the memory 6 and the computing unit 8 are formed in the same layer.

[0145] By adopting Figures 5 to 7 The single package configuration shown may take measures to increase resistance to unauthorized information being obtained from the memory 6 due to disassembly, such as the aforementioned Cu-Cu connection, and may improve tamper resistance in hardware.

[0146] [1-5. Processing procedures]

[0147] Next, we will refer to Figure 8 The flowchart of FIG. 1 describes the processing procedure executed by the calculation unit 8 in order to realize the encryption of the first embodiment described above.

[0148] Furthermore, at least a part of the processing described below can be realized as processing executed by hardware.

[0149] First, as a premise, the computing unit 8 of this example, upon activation and in response to the detection of unauthorized access performed by the unauthorized access detection unit 86, starts Figure 8 The processing shown.

[0150] Initiated in response to detection of unauthorized access Figure 8 The process shown in FIG. 101 acquires an optoelectronic random number and generates an encryption key in response to detection of unauthorized access (S101). Specifically, in response to detection of unauthorized access, the optoelectronic random number is reacquired, and the encryption key is regenerated based on the reacquired optoelectronic random number. Thus, tamper resistance can be improved in software.

[0151] also, Figure 8 The illustrated processing may be started based on other conditions, for example, in response to an external instruction (for example, in response to an instruction of an operation input), or at regular intervals.

[0152] exist Figure 8 In step S101, the calculation unit 8 performs a still image capturing process. This still image capturing process is a process for capturing a still image as a source for generating an encryption key, and the calculation unit 8 controls and causes the array sensor 2 to capture an image of one frame (read out the charge of each pixel).

[0153] Here, one frame of image (still image) which is a source for generating an encryption key is hereinafter referred to as a "seed frame".

[0154] By executing the still image capturing process in step S101 , image data as a seed frame is stored in the memory 6 .

[0155] In step S102 following step S101, the computing unit 8 performs a pixel value uniformity check process. This uniformity check process is a process for checking uniformity between the luminance values ​​of each pixel of the seed frame. Specifically, the computing unit 8 counts the number of pixels whose luminance value is zero or has a saturated value (maximum value).

[0156] Furthermore, the pixel value uniformity check process can also be performed by performing the uniformity check process on the value of the read signal.

[0157] In step S103 following step S102, the calculation unit 8 determines whether the uniformity is excessive. Specifically, it determines whether the number of pixels counted in step S102 is equal to or greater than a predetermined threshold (eg, a value corresponding to 30% to 50% of the number of effective pixels).

[0158] In a case where the number of pixels counted in step S102 is equal to or greater than the above-mentioned threshold value and a determination result of excessive uniformity is obtained, the calculation unit 8 proceeds to step S104 to perform processing of erasing the seed frame, that is, processing of erasing the image data of the seed frame stored in the memory 6, and then returns to step S101.

[0159] Therefore, the seed frame can be re-imaged in response to the situation where the randomness of the pixel values ​​of the seed frame is low. In other words, the photoelectric random number can be re-acquired in response to the situation where the randomness of the photoelectric random number is low.

[0160] Therefore, encryption by an encryption key based on a random number having low randomness can be prevented, and security can be improved.

[0161] On the other hand, in step S103, if the result of determination is obtained that the number of counted pixels is not equal to or greater than the threshold value and the uniformity is not excessive, the calculation unit 8 proceeds to step S105 to generate an encryption key. Specifically, in this example, based on the luminance value of each pixel in the seed frame, an encryption key representing the coefficient set for each amplifier in the amplitude control circuit 10 (or 10A) is generated.

[0162] Here, in the processing of step S105 in this example, an encryption key in a format in which the luminance value of each pixel is directly assigned to the pixel position from which the luminance value is obtained is not generated, but an encryption key in a format in which at least some luminance values ​​of each pixel are assigned to pixel positions different from the pixel positions from which the luminance value is obtained is generated.

[0163] This makes the encryption key difficult to decrypt and can improve security.

[0164] In step S106 following step S105 , the calculation unit 8 performs a process of erasing the seed frame, that is, a process of erasing the image data as the seed frame stored in the memory 6 by the image capturing process of step S101 .

[0165] By performing this seed frame erasing process, it is possible to prevent the image and the estimated photoelectric random number, which are the source of the photoelectric random number, from being leaked.

[0166] For example, if the processing power of the calculation unit 8 is high and the image size of the seed frame is small, it is not necessary to temporarily store the seed frame in the memory 6. In this case, the calculation unit 8 (encryption control unit 85) receives the photoelectric random number from, for example, the amplitude control circuit 10 (or 10A) and generates an encryption key in step S105 through the processing of steps S102 and S103. In this case, the erasure processing of step S106 is unnecessary (of course, the erasure processing of step S104 is also unnecessary).

[0167] In the next step S107, the computing unit 8 erases the existing key (if any). Figure 8 When the illustrated process is started at regular intervals, the encryption key is stored in the memory 6 by the process of step S108 executed in the past. When an existing encryption key is stored in the memory 6, the process of step S107 is a process of erasing the existing encryption key.

[0168] By performing such erasure processing of the existing key, leakage of the encryption key used for encryption in the past can be prevented, and prevention of unauthorized decryption of signals encrypted in the past can be achieved.

[0169] In the next step S108, the calculation unit 8 performs encryption key storage processing. That is, it performs processing to store the encryption key generated in step S105 in the memory 6.

[0170] In response to the execution of the storage process in step S108, the calculation unit 8 completes Figure 8 A series of processing shown.

[0171] In the sensor device 1, the encryption key stored in step S108 is used to encrypt the image signal obtained by imaging using the array sensor 2. Specifically, Figure 8 After the processing shown is completed, the calculation unit 8 (encryption control unit 85) sets the coefficient of each pixel based on the encryption key stored in each amplifier in the amplitude control circuit 10 (or 10A), and performs encryption based on the stored encryption key on the image signal obtained by imaging using the array sensor 2.

[0172] In this example, it is assumed that the array sensor 2 captures a moving image, and encryption performed by the amplitude control circuit 10 (or 10A) is performed on each frame image constituting the moving image.

[0173] Here, as can be understood from the above description, in this example, encryption of the image signal is performed based on the photoelectric random number obtained in a frame period different from that of the encryption target image signal.

[0174] This increases the difficulty of estimating the encryption key from the encrypted image and can improve security.

[0175] Furthermore, the image signal may also be encrypted based on a photoelectric random number obtained in the same frame period as that of the encryption target image signal.

[0176] [1-6. Output information of sensor device]

[0177] Here, the calculation unit 8 may perform an analysis of the properties and actions of the target object, specifically, in this example, an object of the target category from among the above categories.

[0178] For example, in a case where the image obtained by imaging by the array sensor 2 is a surveillance image in a store, by using image analysis, the number, gender, age, height, weight and other attributes of the target customers (people) can be analyzed, whether they wear glasses and whether they wear hats, or the behavior of customers can be analyzed, for example, the movement trajectory (streamline) and changes in posture of customers in the store.

[0179] Alternatively, in the case where the image obtained by imaging by the array sensor 2 is a monitoring image of a vehicle traveling on a road, attributes such as the type and color of the target vehicle, the number of passengers, and the gender of the passengers can be analyzed, or actions such as streamlines and speed on the road can be analyzed.

[0180] like Figure 9 As shown, the sensor device 1 can also adopt a configuration in which data representing the above-mentioned analysis results ("analysis result data" in the figure) is output to the external processor 11 together with motion image data on which encryption according to the above-mentioned embodiment has been performed ("encrypted motion image data" in the figure).

[0181] In this case, the interface unit 7 may output the analysis result data as metadata of the encrypted moving image data, or may output it separately from the encrypted moving image data.

[0182] Furthermore, regarding the analysis result data, it is also conceivable that, for example, the external processor 11 instructs the interface unit 7 on necessary information, and the interface unit 7 outputs the corresponding information.

[0183] Here, in the encryption method of this example, since encryption is performed when the signal is read, it is necessary to decrypt the encrypted image signal in order for the computing unit 8 to perform object detection and the above-mentioned analysis. In this example, the computing unit 8 performs these object detection and analysis processes while decrypting the encrypted image data stored in the memory 6 using the just-in-time method. The just-in-time method is a method of processing encrypted data while decrypting it in units of bytes or words.

[0184] This makes it possible to reduce the possibility of leakage of image signals in a plaintext state when performing object detection from an image or analyzing a detected object, and to improve security.

[0185] [1-7. Regarding key transmission]

[0186] refer to Figure 10 In this article, an example of a method for securely transmitting a key for decryption to a receiving side of an image will be described. This method is a method of transmitting a public key using public key encryption technology.

[0187] First, the receiving side of the image creates a public key and a private key, and transmits the public key to the camera device 100 having the sensor device 1 ( Figure 10 A).

[0188] On the camera device 100 side that has acquired the public key, an encryption key generated based on the photoelectric random number is used as the common key as described above, and the common key is encrypted by the public key and transmitted to the receiving side of the image ( Figure 10 B) in.

[0189] Then, on the receiving side of the image, use Figure 10 A( Figure 10 The private key created in C) in the above example decrypts the transmitted common key (decryption key).

[0190] Therefore, the receiving side of the image can decrypt the encrypted moving image data received from the camera device 100 at high speed by using the decrypted common key later. Figure 10 D in FIG. 1 schematically shows that image encryption is performed on the camera device 100 side using a common key, and Figure 10 E in FIG. 1 schematically shows that the image data encrypted by the common key is decrypted by the receiving side of the image using the public key.

[0191] In addition, despite the Figure 10 The public key encryption common key shown in B and Figure 10 It takes about a few seconds for the private key shown in C to decrypt the common key, but only one encryption and decryption is required for each key transmission.

[0192] <2. Second embodiment>

[0193] [2-1. Encryption method of the second embodiment]

[0194] Next, an encryption method as a second embodiment will be described.

[0195] In addition, in the following description, parts similar to those already explained will be denoted by the same reference numerals, and description thereof will be omitted. In the second embodiment, the configuration of the sensor device 1 is also the same as that of the first embodiment. Figure 1 The same as shown, so their repeated description will be omitted.

[0196] In the second embodiment, a target area of ​​an image signal is encrypted.

[0197] Specifically, in the second embodiment, the entire image and the target area are encrypted based on different encryption keys, and the target area is encrypted based on different encryption keys in a specific part area and other areas, thereby gradually changing the hiding level of information according to the decryption key stored on the image receiving side.

[0198] Figure 11 and Figure 12 : is a diagram for explaining the progressively encrypted image in the second embodiment.

[0199] Figure 11 Shown are images that are progressively encrypted when the target class is a person.

[0200] Figure 11 A in FIG shows an image before encryption. In this case, the target area AT is the entire area where the person is projected in the image. In addition, in this case, the specific area AS, which is the area of ​​the specific portion, is the face area of ​​the person.

[0201] Figure 11 B in FIG. 1 shows an image in which only a specific area AS is encrypted. Figure 11 The specification C shows that only the image of the target area AT including the specific area AS is encrypted, Figure 11 D in FIG. 1 shows an image in which the entire image is encrypted.

[0202] Figure 12 shows the progressively encrypted images where the target class is a vehicle, and Figure 12 A shows the image before encryption.

[0203] In this case, the target area AT is the entire area in the image where the vehicle is projected, and the specific area AS includes the areas of the vehicle's passengers and license plate.

[0204] Figure 12 B in the image is an image in which only a specific area AS is encrypted. Figure 12C in FIG. 1 is an encrypted image including only the target area AT of the specific area AS, and Figure 12 The D in is the image where the entire image is encrypted.

[0205] Here, Figure 12 The image shown in Figure A before encryption is individually identifiable personal information because it shows the passenger and license plate. Therefore, it is not suitable for marketing data.

[0206] When the image is only Figure 12 When encrypted in a specific area shown in B in the image, for example, for marketing purposes of a vehicle manufacturer, the type and number of vehicles can be determined without obtaining personally identifiable information, such as passengers including the driver and license plate. Figure 12 When the target area AT shown in C in FIG. 1 is selected, only information about the number and movement of vehicles can be obtained without obtaining personal information or vehicle type information. For example, congestion conditions can be determined.

[0207] exist Figure 11 and Figure 12 In the example case, at least three types of encryption keys are generated as encryption keys: a first encryption key supporting encryption of the entire image, a second encryption key supporting encryption of only the target area AT, and a third encryption key supporting encryption of only the specific area AS.

[0208] Figure 13 is a diagram for explaining an example of a specific method of stepwise encryption.

[0209] In this example, encryption based on the first, second, and third encryption keys is not performed on the target image individually, but encryption is performed on the target image based on a synthesized key obtained by synthesizing the plurality of encryption keys.

[0210] First, a third encryption key for encrypting the specific area AT, a second encryption key for encrypting the entire target area AT, and a first encryption key for encrypting the entire image are prepared. When generating these three types of encryption keys, three types of photoelectric random numbers can be obtained (i.e., three types of seed frames are imaged). However, in this example, to shorten the time required to generate the encryption keys, the three types of encryption keys are generated from a common photoelectric random number. Specifically, in this example, when generating the three types of encryption keys, three types of random numbers (hereinafter referred to as the first random number, the second random number, and the third random number, respectively) are first generated, where the arrangement of the numerical values ​​for each pixel in the common photoelectric random number is different.

[0211] Then, the third encryption key is generated as an encryption key obtained by extracting the numerical value of each pixel of the specific area AS from the numerical value of the third random number.

[0212] Furthermore, the second encryption key is generated as an encryption key obtained by extracting a numerical value of each pixel of the target area AT from the numerical value of the second random number.

[0213] The first encryption key is generated by directly applying the encryption key generated by the first random number.

[0214] Then, as shown in the figure, an encryption key obtained by synthesizing these first, second, and third encryption keys is generated as a synthesized key.

[0215] Then, based on the synthesized key, the target image is encrypted.

[0216] By performing the above-described stepwise encryption, the concealment level of information can be changed according to the decryption key held by the receiving side of the image.

[0217] Figure 14 It is an explanatory diagram of an example of a change in the hidden level.

[0218] Here, four levels related to key storage are defined, from Level 0 to Level 3. As shown in the figure, Level 0 indicates no key, Level 1 indicates holding only the first encryption key, Level 2 indicates holding a composite key of the first and second encryption keys, and Level 3 indicates holding a composite key of the first, second, and third encryption keys.

[0219] In the case of level 0, the encrypted image cannot be decrypted on the receiving side of the image, and an image in which the entire area is encrypted can be obtained.

[0220] In the case of Level 1, the reception side of the image can decrypt the area other than the target area AT by using the first encryption key, and thus can obtain an image in which only the target area AT is encrypted.

[0221] In the case of Level 2, the reception side of the image can decrypt the area other than the specific area AS by using the synthetic key of the first and second encryption keys, and thus can obtain an image of only the specific area AS in the encryption target.

[0222] In the case of Level 3, the receiving side of the image can decrypt the entire image using a composite key of the first, second, and third encryption keys, and in this case, an image in which information is not hidden can be obtained.

[0223] Here, in this example, since the encryption target image is a moving image, the object as the target projected in the image may shift in the image over time. For this reason, when encryption is performed on the target area AT as described above, it is necessary to track the target.

[0224] In the following, reference will be made to Figure 15An example of a specific method of encryption as a second embodiment including tracking of such a target is described.

[0225] In addition, Figure 15 In , the target class is assumed to be “person”. In addition, Figure 15 In the foregoing, for convenience of explanation, an example is given in which the specific area AC and other areas are not distinguished in the target area AT, and only the target area AT and other areas in the image are encrypted separately.

[0226] first, Figure 15 Frame F1 shown in A in FIG shows a state where a person in the target category is not yet in the frame. Here, an example is given of identifying an object that is not a "tree" in the target category in an image.

[0227] Here, the entire image is encrypted regardless of the presence or absence of the target. That is, in this example, the amplitude control circuit 10 (or 10A) encrypts the image of each frame F based on the first encryption key corresponding to the entire image and then stores it in the memory 6. Figure 15 The white key mark shown in each sub-figure of indicates that encryption has been performed on the entire image as the output image.

[0228] To track the target, the computing unit 8 detects an object region in the image or identifies its class (processing by the object region identification unit 82 and the class identification unit 83 described above). To perform these processes, the computing unit 8 decrypts the frame image encrypted and stored as described above. In other words, the computing unit 8 performs the target tracking process while decrypting the frame image encrypted using the first encryption key.

[0229] The calculation unit 8 performs decryption using the on-the-fly method at this time. As a result, the possibility of leakage of the plaintext image signal when performing target tracking can be reduced, and security can be improved.

[0230] Figure 15 Frame F2 shown in B in FIG shows the state of the target class of "person" in the frame. In this state, the target class "person" is recognized together with the already recognized "tree".

[0231] When an object of the target class is identified in this manner, the calculation unit 8 (the object region identification unit 82 ) performs calculation of the bounding box 20 using accurate position coordinates around the object region.

[0232] For example, Figure 15 C in FIG. 1 shows an example of a bounding box 20 of an image of a person as a target class. That is, the bounding box 20 is calculated as a more accurate region corresponding to an object of the target class.

[0233] Furthermore, the calculation unit 8 (the object region identification unit 82 ) calculates a ROI 21 as a region of interest based on the bounding box 20 .

[0234] Figure 15 D in FIG. 1 shows the ROI 21 and the bounding box 20. The ROI 21 is calculated by, for example, enlarging (ax×by) the vertical and horizontal dimensions (x×y) of the bounding box 20. Although the enlargement ratios a and b can be set for the vertical and horizontal directions, respectively, and the enlargement ratios thereof can be fixed, it is also conceivable that the sensor device 1 is specified from the outside (e.g., the external processor 11, etc.).

[0235] In this example, the ROI 21 is set as the target area AT, and encryption is performed using an encryption key different from that of the entire image.

[0236] Here, the frame F2 is a frame in which an object class is newly recognized in an image, in other words, it can be called an object class discovery frame.

[0237] In this example, since a method of encrypting the read signal from the pixel is adopted, encryption based on the second encryption key cannot be performed on the ROI 21 in the target class discovery frame. The target class discovery frame has been encrypted based only on the first encryption key to be stored in the memory 6. In this way, if the target class discovery frame that has been encrypted based only on the first encryption key is output as is, the image area of ​​the ROI 21 will be disclosed only to the holder of the first encryption key without being hidden.

[0238] Therefore, in this example, the target class discovery frame is erased from the memory 6, thereby achieving an appropriate level of information hiding according to the decryption key held by the receiving side of the image.

[0239] Figure 15 E in FIG. 5 represents frame F3, which is a frame following frame F2.

[0240] From the frame F next to the target class discovery frame, encryption based on the second encryption key is performed for the ROI 21. The ROI 21 here is the ROI 21 calculated when the frame F2 is the target class discovery frame.

[0241] In the case where a "person" as the target class is moving, the person in frame F3 moves further from frame F2 in the direction of movement, but by setting ROI 21 to have a larger range than bounding box 20, the person as the target class can be accommodated in ROI 21 in frame F3. In other words, the person as the target class falls within the target range encrypted based on the second encryption key.

[0242] Similarly, after frame F3, the bounding box 20 and ROI 21 of the target class are also calculated, thereby tracking the target class (see Figure 15 F in.

[0243] In addition, after the frame F4, as in the frame F3, the ROI 21 calculated in the previous frame F is encrypted based on the second encryption key (see Figure 15 G in ).

[0244] Figure 15 H in FIG shows a frame Fn after the target class "person" is outside the frame. ROI 21 is no longer calculated because the target class is not within the frame. Therefore, encryption based only on the first encryption key is performed only on the image of frame Fn.

[0245] In the above description, an example has been described in which a rectangular region obtained by enlarging the bounding box 20 is set as the ROI 21 , but the ROI 21 is not limited to a rectangular region.

[0246] The ROI 21 may be calculated from regions of objects of the target class using, for example, semantic segmentation, ie, pixel-level object region detection.

[0247] Figure 16 ROI 21 based on semantic segmentation is shown. This is an example of setting a non-rectangular ROI 21 by expanding a pixel area that is an object (eg, a person).

[0248] The rectangular ROI 21 may not include a portion or may be too large in the case of, for example, a truck with a protrusion, a cyclist, etc. When the non-rectangular ROI 21 is generated based on the position of the object position at the pixel level, the hidden area related to the target can be appropriately set without excess or deficiency.

[0249] [2-2. Processing Procedure]

[0250] Will refer to Figure 17 and Figure 18 The flowchart of FIG. 1 describes the processing procedure executed by the calculation unit 8 in order to realize the encryption as described above in the second embodiment.

[0251] Figure 17 The process corresponding to the process from imaging the seed frame to storing the random number as the encryption key source is shown. Figure 17 , the same step number is assigned to the step that is already in Figure 8 The same processing as described in , and its description will be omitted.

[0252] Similar to Figure 8 processing, Figure 17The processing starts upon activation and in response to detection of unauthorized access performed by the unauthorized access detection unit 86. Alternatively, the processing may be started based on other conditions, for example, at regular time intervals.

[0253] In addition, reference Figure 17 and Figure 18 At least a portion of the described processing can be implemented as a process using hardware.

[0254] exist Figure 17 In this case, if the calculation unit 8 determines in step S103 that the uniformity is excessive, the calculation unit 8 proceeds to step S201 to generate a random number for each level. Here, since the specific area AS is not distinguished in the target area AT, two types of random numbers are generated, namely the first random number and the second random number described above.

[0255] Furthermore, since the method of generating various random numbers based on the photoelectric random number of the seed frame has been described, repeated description thereof will be omitted.

[0256] In this case, in response to the random number generation process being performed in step S201 , the calculation unit 8 performs the seed frame erasure process in step S106 .

[0257] Then, in response to the execution of the erasure process in step S106, the calculation unit 8 executes a process of erasing any existing random numbers in step S202. That is, in the process of step S203 executed in the past, when there are random numbers (first random numbers and second random numbers) for each level stored in the memory 6, it is a process of erasing the random numbers.

[0258] In step S203 following step S202, the calculation unit 8 performs a process of storing the random number of each level generated in step S201 in the memory 6, and ends the process. Figure 17 A series of processing shown.

[0259] Figure 18 A process for encrypting a target image based on a generated encryption key is shown.

[0260] First, in step S301, computing unit 8 waits for the start of imaging of the encryption target image. When imaging begins, computing unit 8 executes encryption processing using the first encryption key in step S302. Specifically, the coefficient for each pixel is instructed to amplitude control circuit 10 (or 10A) based on the first encryption key, and encryption of the read signal from array sensor 2 is executed. As will be understood from the above description, in this example, the first encryption key is an encryption key to which the first random number is applied as is.

[0261] In step S303 following step S302, the calculation unit 8 performs object region recognition processing, and performs category recognition processing in the subsequent step S304. The object region recognition processing in step S303 is the processing of the above-mentioned object region recognition unit 82, and performs processing for detecting candidate objects from the image of the current frame and recognizing their object regions. In addition, the category recognition processing in step S304 is the processing of the above-mentioned category recognition unit 83, and category recognition is performed on the objects detected in the above-mentioned object region recognition processing. In the case where multiple objects or multiple types of objects are detected, category recognition is performed on each object, and they are classified into each category. For example, in the above-mentioned Figure 15 In the case of B, for example, category recognition and classification are performed so that there is one object of the category "tree" and one object of the category "person".

[0262] Furthermore, the calculation unit 8 performs the processing of steps S303 and S304 while decrypting the frame image encrypted in step S302 and step S313 described later using the on-the-fly method.

[0263] In step S305 following step S304, the calculation unit 8 determines whether the target class exists. That is, it determines whether the target class exists among the classes identified in step S304.

[0264] If no target class exists, the computing unit 8 waits for the next frame (wait for the arrival of the next frame period) in step S306 and then returns to step S302. That is, until a target class is detected, the encryption process of the entire image in step S302, the object region identification process in step S303, and the class identification process in step S304 are repeatedly executed frame by frame.

[0265] In the case where it is determined in step S305 that the target class exists, the calculation unit 8 proceeds to step S307 to calculate the bounding box 20 and then calculates the ROI 21 in step S308 .

[0266] Furthermore, in the next step S309 , the calculation unit 8 generates a composite key in which the second encryption key and the first encryption key are combined by applying the value of the second random number only to the ROI 21 .

[0267] In response to the generation of the synthetic key in step S309, the computing unit 8 determines in step S310 whether it is a target class discovery frame. If the current frame is a target class discovery frame, the computing unit 8 performs a process of erasing the target class discovery frame in step S311. As a result, even if the key retention level is level 1, the target image portion of the target class discovery frame can be prevented from being concealed.

[0268] In step S310, if the current frame is not a target class discovery frame, the calculation unit 8 skips the erasure process of step S311 and performs a process of waiting for the next frame in step S312. In addition, when performing the erasure process of step S311, the calculation unit 8 also performs a process of waiting for the next frame in step S312.

[0269] In response to the execution of the waiting process in step S312, the calculation unit 8 performs encryption processing in step S313 using the synthetic key generated in the previous frame. That is, the coefficient of each pixel of the amplitude control circuit 10 (or 10A) is instructed based on the synthetic key to encrypt the read signal of the array sensor 2.

[0270] In step S314 following step S313 , the calculation unit 8 determines whether imaging is completed, that is, whether imaging of the encryption target image should be terminated, for example, when an external imaging end instruction is given, or the like.

[0271] If imaging is not complete, the computing unit 8 returns to step S303. Thus, the processing described so far is repeated until imaging is complete. Specifically, if the target class continues to exist, calculation of the ROI for the target class, generation of a composite key based on the calculated ROI, and encryption processing based on the composite key generated in the previous frame are performed. If the target class no longer exists, encryption processing using the composite key is not performed, and encryption processing using the first encryption key is performed.

[0272] When imaging is completed, the calculation unit 8 ends Figure 18 A series of processing shown.

[0273] Furthermore, in this example, the ROI 21 is set as an extended area of ​​the bounding box 20 so as to be able to include the target object in the next frame, but it is also conceivable that when the vertical and horizontal sizes (x×y) are enlarged (ax×by), the enlargement ratios a and b correspond to the frame rate.

[0274] For example, if the frame rate is low, the frame interval time becomes longer and the movement amount of an object (eg, a person) increases, so it is conceivable to widen the ROI 21 compared to the case where the frame rate is high.

[0275] Furthermore, in the case of performing encryption on the target area AT by distinguishing a specific portion from other areas, the bounding box 20 and the ROI 21 are calculated for the specific portion using the same method as described above, and a third encryption key is generated by applying the value of the third random number to the calculated ROI 21. Then, a composite key obtained by synthesizing the first, second, and third encryption keys can be generated and used to encrypt the image of the next frame.

[0276] [2-3. Example of output of analysis information]

[0277] Furthermore, in the second embodiment, the sensor device 1 may also be configured to output the analysis result data together with the encrypted moving image data (see Figure 9 ).

[0278] At this time, in the second embodiment, since encryption is performed for the target area AT, the analysis result data is not limited to text data, and may be data that can be displayed as an image, for example, as shown in FIG. Figure 19 shown.

[0279] exist Figure 19 middle, Figure 19 A in represents the original image before encryption.

[0280] Figure 19 B in FIG. 1 is a diagram showing an example of displaying attribute information analyzed for a target (here, “person”) on an image, in which gender information of the target is distinguishably displayed by a color added to an encrypted target area.

[0281] also, Figure 19 C in FIG. 1 is an example in which the posture information Ib obtained by the analysis processing of the target is superimposed on the target area and displayed in an encrypted state.

[0282] By performing encryption for the target area in this manner, it is possible to provide an image that makes it easier to understand the analysis results of the attributes and behavior of the target while ensuring the confidentiality of personal information.

[0283] In recent years, in-store camera images have increasingly been used as marketing data. In these cases, identifying individual customers is not necessary, but rather analyzing their attributes and behavior. Figure 19 The image in A is individually identifiable data and is highly confidential information, so a system with high security measures is required, resulting in high costs. Figure 19 The image shown in B in FIG is useful as marketing data because the flow of customers in the store and the attributes of men and women can be seen. Figure 19 The image shown in center C further recognizes the customer's posture, so the behavior of customers who pick up products or try them on but do not purchase can be determined, which can be used as high-value marketing data.

[0284] Because in Figure 19 The image of B and Figure 19 No individual can be identified in the images of C, so Figure 19Compared to the image in Figure A, this has the advantage of being operable within a low-cost security system. Furthermore, using a camera equipped with the sensor device 1 of this embodiment means that even if malware is installed on the in-store camera, there is no plaintext data, and there is no risk of hacker attacks.

[0285] <3. Modifications>

[0286] Furthermore, the embodiments are not limited to the specific examples described so far, and various modifications can be considered.

[0287] For example, although not specifically mentioned above, the present technology can also be appropriately applied to a case where the array sensor 2 having a color filter based on, for example, the Bayer array is used and a color image is obtained as a captured image.

[0288] When using an array sensor 2 with color filters, for example, a predetermined number of pixels consisting of vertical lines × horizontal lines = multiple pixels × multiple pixels are considered as one color unit, and pixel values ​​are combined for each color unit to obtain a color signal value, such as an RGB value. For example, if a Bayer array is used, vertical lines × horizontal lines = 2 × 2 = 4 pixels on which RGGB color filters are formed are considered as one color unit, and each pixel value (luminance value) of RGGB is combined for each color unit to obtain a set of RGB values.

[0289] When the array sensor 2 having a color filter is used, an encryption key in which a random value is assigned to each color cell can also be generated as an encryption key for image encryption.

[0290] Figure 20 is an explanatory diagram of an example of generating an encryption key, in which a random value is assigned to each color cell.

[0291] exist Figure 20 In the example of A in FIG, the brightness value of a pixel in a color unit is assigned a random value for each pixel in the unit. Specifically, in the example of the figure, the brightness value of the upper left corner pixel is assigned a random value for each pixel in the unit of each color unit.

[0292] exist Figure 20 In the example of B in FIG, a value calculated by a predetermined calculation using the brightness values ​​of the pixels in the cell is assigned as a random value for each pixel in each color cell. As a specific example, it is conceivable that the average value of the brightness values ​​of each pixel in the cell is assigned as a random value for each pixel in the unit of each color cell. That is, with respect to the color cell located in the upper left corner of the figure, Vmix1 = (V1 + V2 + V7 + V8) / 4.

[0293] In this case, calculations using the luminance values ​​of all pixels in a cell are not necessarily performed, and calculations can be performed using only the luminance values ​​of some pixels. For example, it can be set to Vmix1 = (V1 + V7) / 2, etc. In addition, the average value is not necessarily used for each cell. For example, it is conceivable to assign the total value of the luminance value of each pixel to each cell, for example, Vmix1 = V1 + V2 + V7 + V8.

[0294] As described above, by generating an encryption key that assigns a random value to each color cell, the processing load can be reduced compared to the case where a random value is assigned to each pixel.

[0295] Furthermore, in the above, an example has been described in which the encryption target signal is an image signal encrypted using photoelectric random numbers, but the encryption target signal is not limited to the image signal.

[0296] While the above examples describe methods for encrypting signals read from pixels of the array sensor 2 and methods for encrypting a target area using photoelectric random numbers, the random numbers used for encryption in these methods are not limited to photoelectric random numbers. For example, pseudorandom numbers can be used. Alternatively, in the case of using true random numbers, for example, a method can be mentioned in which a sensor detects natural phenomena (e.g., changes in heat or sound, which are practically impossible to predict or reproduce) and generates random numbers based on these values.

[0297] <4. Overview of Embodiments>

[0298] As described above, as described in the first embodiment, the encryption device (sensor device 1) of this embodiment includes: an encryption key generation unit (encryption control unit 85) that generates an encryption key based on a photoelectric random number, wherein the photoelectric random number is a random number obtained based on the photoelectric conversion performed by the array sensor (array sensor 2), in which a plurality of pixels each having a light receiving element for visible light or invisible light are arranged one-dimensionally or two-dimensionally; and an encryption unit (amplitude control circuit 10 or 10A) that encrypts the target signal based on the encryption key generated by the encryption key generation unit.

[0299] Therefore, encryption of the encryption key that is more difficult to decrypt than the case of using pseudo-random numbers can be achieved.

[0300] Therefore, security can be improved.

[0301] Furthermore, in the encryption device of this embodiment, the encryption key generation unit acquires the value of the electric signal of each pixel obtained by photoelectric conversion as a photoelectric random number to generate the encryption key.

[0302] Therefore, an encryption key that is difficult to decrypt can be generated.

[0303] Therefore, security can be improved.

[0304] Furthermore, in the encryption device of this embodiment, the encryption unit performs encryption on the image signal obtained by imaging using the array sensor based on the encryption key.

[0305] Therefore, encryption can be performed on the image signal using the encryption key, and the coefficient used for encryption is allocated to each pixel of the array sensor.

[0306] Therefore, there is no need to perform complicated calculation processing for encrypting the image signal, and the encryption processing speed can be increased.

[0307] Furthermore, in the encryption device of this embodiment, the encryption key generation unit generates an encryption key in a format in which at least some values ​​of the electric signal of each pixel are assigned to pixel positions different from pixel positions from which the electric signal values ​​are obtained.

[0308] Therefore, it is difficult to decrypt an encryption key in which the electric signal value of each pixel is directly assigned to the pixel position from which the electric signal value is obtained, compared to the case of using an encryption key.

[0309] Therefore, security can be improved.

[0310] Furthermore, in the encryption device of this embodiment, the encryption key generation unit generates the encryption key based on the photoelectric random number obtained in a frame period different from a frame period of the image signal to be encrypted by the encryption unit.

[0311] Therefore, the difficulty of estimating the encryption key from the encrypted image increases.

[0312] Therefore, security can be improved.

[0313] Furthermore, in the encryption device of this embodiment, in the case where the uniformity of the electric signal value has been identified in at least some of the pixels, the encryption key generation unit reacquires the photoelectric random number.

[0314] Therefore, encryption can be prevented from being performed by an encryption key based on a random number having low randomness.

[0315] Therefore, security can be improved.

[0316] Furthermore, in the encryption device of the present embodiment, the encryption key generation unit, the encryption unit, and the array sensor are configured in one package.

[0317] Therefore, tamper resistance can be improved in hardware.

[0318] Therefore, security can be improved.

[0319] Furthermore, in the encryption device of this embodiment, the encryption key generation unit reacquires the photoelectric random number in response to detection of unauthorized access from outside the encryption device.

[0320] Therefore, after unauthorized access from the outside is detected, encryption can be performed based on the re-acquired optoelectronic random number.

[0321] Therefore, security can be improved.

[0322] Furthermore, in the encryption device of this embodiment, in response to reacquisition of the photoelectric random number, the encryption key generation unit erases the previously generated encryption key from the memory.

[0323] Therefore, leakage of the encryption key used for encryption in the past can be prevented.

[0324] Therefore, unauthorized decryption of a signal encrypted in the past can be prevented, and security can be improved.

[0325] Furthermore, in the encryption device of this embodiment, the encryption key generation unit erases the image signal, which is the source of the photoelectric random number, from the memory in response to the generation of the encryption key.

[0326] Therefore, it is possible to prevent the image which is the source of the photoelectric random number and the estimated photoelectric random number from being leaked.

[0327] Therefore, security can be improved.

[0328] Furthermore, in the encryption device of this embodiment, the encryption unit encrypts the target signal using the stream cipher method.

[0329] Therefore, no pre-processing of the encryption of the target signal is required.

[0330] Therefore, the encryption process can be accelerated.

[0331] In addition, the encryption method of this embodiment is an encryption method, including: generating an encryption key based on a photoelectric random number, which is a random number obtained based on the photoelectric conversion performed by an array sensor, in which a plurality of pixels each having a light receiving element for visible light or invisible light are arranged one-dimensionally or two-dimensionally; and performing encryption of a target signal based on the generated encryption key.

[0332] Using this encryption method, the same operation and effects as those of the encryption device of the above-mentioned embodiment can also be obtained.

[0333] In addition, as described in the first embodiment, the sensor device (sensor device 1) of this embodiment includes an array sensor (array sensor 2) and an encryption unit (amplitude control circuit 10 or 10A, or encryption control unit 85), in which a plurality of pixels of light receiving elements for visible light or invisible light are arranged one-dimensionally or two-dimensionally, and the encryption unit encrypts the read signals from the pixels of the array sensor.

[0334] By performing encryption of the read signal in this manner, it is possible to prevent the plaintext image signal from being stored in the memory.

[0335] Therefore, security can be improved.

[0336] Furthermore, in the sensor device of the present embodiment, the encryption unit has a first amplitude control unit (amplitude control circuit 10) that performs amplitude control of a read signal that is an analog signal, and performs encryption of the read signal by performing amplitude control based on an encryption key in the first amplitude control unit.

[0337] It is difficult to obtain a read signal as an analog signal from outside the sensor device.

[0338] Therefore, security can be improved.

[0339] Furthermore, in the sensor device of this embodiment, the encryption unit has a second amplitude control unit (amplitude control circuit 10A) that performs amplitude control on the read signal converted into a digital signal by the A / D converter, and encrypts the read signal by performing amplitude control based on an encryption key in the second amplitude control unit.

[0340] Therefore, encryption is performed as amplitude control on a digital signal, and the accuracy of encryption processing is improved compared to the case where amplitude control on an analog signal is performed.

[0341] Therefore, when the encrypted image is decrypted, the reproducibility of the image content can be improved.

[0342] Furthermore, in the sensor device of the present embodiment, the array sensor and the encryption unit are configured in one package.

[0343] Therefore, tamper resistance can be improved in hardware.

[0344] Therefore, security can be improved.

[0345] Furthermore, in the sensor device of this embodiment, the encryption unit generates an encryption key based on a photoelectric random number which is a random number obtained based on photoelectric conversion performed by the array sensor, and encrypts the read signal based on the generated encryption key.

[0346] Therefore, encryption of the encryption key that is more difficult to decrypt than the case of using pseudo-random numbers can be achieved.

[0347] Therefore, security can be improved.

[0348] Furthermore, in the sensor device of this embodiment, the encryption unit generates the encryption key based on the photoelectric random number obtained in a frame period different from a frame period of the encryption target read signal.

[0349] Therefore, the difficulty of estimating the encryption key from the encrypted image increases.

[0350] Therefore, security can be improved.

[0351] Furthermore, in the sensor device of this embodiment, the encryption unit reacquires the photoelectric random number in response to detection of unauthorized access from outside the sensor device.

[0352] Therefore, after unauthorized access from the outside is detected, encryption can be performed based on the re-acquired optoelectronic random number.

[0353] Therefore, security can be improved.

[0354] Furthermore, in the sensor device of this embodiment, the encryption unit erases the encryption key generated in the past from the memory in response to reacquisition of the photoelectric random number.

[0355] Therefore, leakage of the photoelectric random number used for encryption in the past can be prevented.

[0356] Therefore, unauthorized decryption of a signal encrypted in the past can be prevented, and security can be improved.

[0357] Furthermore, in the sensor device of this embodiment, the encryption unit erases the image signal, which is the source of the photoelectric random number, from the memory in response to the generation of the encryption key.

[0358] Therefore, it is possible to prevent the image which is the source of the photoelectric random number and the estimated photoelectric random number from being leaked.

[0359] Therefore, security can be improved.

[0360] Furthermore, another encryption method of the embodiment is an encryption method of performing encryption on read signals from pixels of an array sensor in which a plurality of pixels each having a light receiving element for visible light or invisible light are arranged one-dimensionally or two-dimensionally.

[0361] Using this encryption method, the same operation and effects as those of the sensor device of the above-described embodiment can be obtained.

[0362] In addition, as described in the second embodiment, another sensor device (sensor device 1) of the present embodiment includes: an array sensor (array sensor 2) in which a plurality of pixels, each having a light receiving element for visible light or invisible light, are arranged one-dimensionally or two-dimensionally; a detection unit (computing unit 8: in particular, an object area identification unit 82 or a category identification unit 83) that detects an area of ​​a target projected in an image as a target area based on an image signal obtained by imaging using the array sensor; and an encryption unit (amplitude control circuit 10 or 10A, or encryption control unit 85) that performs encryption for the target area in the image signal based on information of the target area detected by the detection unit.

[0363] Therefore, it is possible to perform encryption on an image signal obtained by imaging using an array sensor in an image sensor, at least to the extent that a person cannot be identified.

[0364] Therefore, there is no need to take measures to prevent leakage of personal information on the image receiving side, and costs can be reduced.

[0365] Furthermore, by retaining the decryption key on the receiving side of the image, it is possible to visually identify a portion of the image content while concealing personal information. This allows image encryption to prevent excessive concealment of information.

[0366] For example, excessive information hiding can be prevented, which makes it difficult to grasp even the shooting location, for example, at which store the image was shot, and hiding of personal information and prevention of excessive damage to the usefulness of the image can be achieved.

[0367] Furthermore, in the sensor device of this embodiment, as encryption of the image signal, the encryption unit encrypts the entire image based on the first encryption key, and encrypts the target area based on the first encryption key and a second encryption key different from the first encryption key.

[0368] Therefore, the portion of the image that can be decrypted is differentiated based on the type of decryption key stored on the image receiving side. Specifically, if only the decryption key corresponding to the first encryption key is stored, only the image area excluding the target area can be decrypted. However, if the decryption keys corresponding to both the first and second encryption keys are stored, the entire image including the target area can be decrypted.

[0369] Therefore, a useful encryption method can be realized in which the hiding level of information can be gradually changed according to the decryption key stored on the image receiving side.

[0370] Furthermore, in the sensor device of the present embodiment, the detection unit performs a process of recognizing a specific portion of a target, and performs encryption based on different encryption keys for the region of the specific portion and other regions in the target region.

[0371] Therefore, the concealment level of the target can be changed according to the decryption key stored on the image receiving side. For example, in the case where the target is a person, the concealment level can be classified into, for example, concealment of the whole body and concealment of only its face.

[0372] Therefore, it is possible to provide an encrypted image with an appropriate hiding level according to the usage pattern of the image.

[0373] Furthermore, the sensor device of the present embodiment includes an analyzing unit (computing unit 8 ) that analyzes the attribute or action of a target and an output unit (interface unit 7 ) that outputs information indicating the result of the analysis performed by the analyzing unit.

[0374] Therefore, even if the receiving side of the image does not have a decryption key, analysis results regarding the attributes and actions of the target can be provided.

[0375] Furthermore, in the sensor device of this embodiment, the detection unit detects the target by decrypting the image signal encrypted based on the first encryption key using an on-the-fly method.

[0376] Therefore, the possibility of leakage of the image signal in the plain text state can be reduced.

[0377] Therefore, security can be improved.

[0378] Furthermore, in the sensor device of this embodiment, the encryption unit performs encryption on the image signal based on a synthetic key obtained by synthesizing a plurality of encryption keys.

[0379] Therefore, when encryption is implemented that gradually changes the concealment level of personal information, the number of execution times of the required encryption processing can be reduced.

[0380] Therefore, the processing load associated with encryption can be reduced.

[0381] Furthermore, in the sensor device of this embodiment, the encryption unit performs encryption for the target area while tracking the target.

[0382] Therefore, in the case where the encryption target image is a moving image, the moving object can be appropriately concealed.

[0383] Furthermore, in the sensor device of this embodiment, the encryption unit performs encryption of the image signal using a stream cipher method.

[0384] Therefore, no pre-processing of image signal encryption is required.

[0385] Therefore, the encryption process can be accelerated.

[0386] Furthermore, in the sensor device of the present embodiment, the array sensor, the detection unit, and the encryption unit are configured in one package.

[0387] Therefore, tamper resistance can be improved in hardware.

[0388] Therefore, security can be improved.

[0389] Furthermore, in the sensor device of this embodiment, the encryption unit generates an encryption key based on a photoelectric random number obtained based on photoelectric conversion performed by the array sensor, and performs encryption on the image signal based on the generated encryption key.

[0390] Therefore, encryption of the encryption key that is more difficult to decrypt than the case of using pseudo-random numbers can be achieved.

[0391] Therefore, security can be improved.

[0392] Furthermore, in the sensor device of this embodiment, the encryption unit performs encryption on the read signals from the pixels of the array sensor.

[0393] Therefore, it is possible to prevent the plaintext image signal from being stored in the memory while being encrypted.

[0394] Therefore, security can be improved.

[0395] In addition, another encryption method of this embodiment is an encryption method, including: detecting an area of ​​a target projected in an image based on an image signal obtained by imaging with an array sensor, as a target area, in the array sensor, a plurality of pixels each having a light receiving element for visible light or invisible light are arranged one-dimensionally or two-dimensionally; and based on information of the detected target area, performing encryption for the target area in the image signal.

[0396] Using this encryption method, the same operation and effects as those of the sensor device of the above-described embodiment can be obtained.

[0397] Furthermore, the effects described in this specification are merely examples and are not intended to be limiting, and other effects may be obtained.

[0398] <5. This technology>

[0399] The present technology may also have the following configurations.

[0400] (1) A sensor device comprising:

[0401] An array sensor in which a plurality of pixels each having a light-receiving element for visible or invisible light are arranged one-dimensionally or two-dimensionally;

[0402] a detection unit that detects, as a target area, a region of a target projected in an image based on an image signal obtained by imaging with the array sensor; and

[0403] The encryption unit performs encryption on the target area in the image signal based on the information of the target area detected by the detection unit.

[0404] (2) The sensor device according to (1) above, wherein, in encryption of the image signal, the encryption unit encrypts the entire image based on a first encryption key, and encrypts the target area based on the first encryption key and a second encryption key different from the first encryption key.

[0405] (3) The sensor device according to (2) above,

[0406] wherein the detection unit performs processing for identifying a specific portion of the target, and

[0407] The encryption unit performs encryption based on different encryption keys for an area of ​​a specific portion in a target area and other areas.

[0408] (4) The sensor device according to any one of (1) to (3) above, further comprising:

[0409] Analysis unit, which analyzes the properties or actions of the target, and

[0410] The output unit outputs information indicating the analysis result performed by the analysis unit.

[0411] (5) The sensor device according to any one of (2) to (4) above, wherein the detection unit performs detection of the target by decrypting the image signal encrypted based on the first encryption key using a just-in-time method.

[0412] (6) The sensor device according to any one of (2) to (5) above, wherein the encryption unit encrypts the image signal based on a composite key obtained by synthesizing a plurality of encryption keys.

[0413] (7) The sensor device according to any one of (1) to (6) above, wherein the encryption unit performs encryption for the target area while tracking the target.

[0414] (8) The sensor device according to any one of (1) to (7) above, wherein the encryption unit encrypts the image signal using a stream cipher.

[0415] (9) The sensor device according to any one of (1) to (8) above, wherein the array sensor, the detection unit, and the encryption unit are configured in one package.

[0416] (10) A sensor device according to any one of (1) to (9) above, wherein the encryption unit generates an encryption key based on a photoelectric random number, and performs encryption of the image signal based on the generated encryption key, wherein the photoelectric random number is a random number obtained based on the photoelectric conversion performed by the array sensor.

[0417] (11) The sensor device according to any one of (1) to (10) above, wherein the encryption unit encrypts a read signal from a pixel of the array sensor.

[0418]

Explanation of symbols

[0419] 1 sensor device

[0420] 2 array sensors

[0421] 3ADC / Pixel Selector

[0422] 4 buffers

[0423] 5 logic units

[0424] 6 Memory

[0425] 7Interface Unit

[0426] 8 computing units

[0427] 10. 10A amplitude control circuit

[0428] 20 bounding boxes

[0429] 21ROI

[0430] 82 target area recognition unit

[0431] 83 Category Identification Unit

[0432] 85 encryption control unit

[0433] 86 Unauthorized access detection unit

[0434] 100 camera setup.

Claims

1. A sensor device, comprising: An array sensor in which a plurality of pixels each having a light receiving element for visible light or invisible light are arranged one-dimensionally or two-dimensionally; a detection unit that detects a region of a target projected in an image as a target region based on an image signal obtained by imaging with the array sensor; and an encryption unit that performs encryption for the target area in the image signal based on information of the target area detected by the detection unit, In the encryption of the image signal, the encryption unit encrypts the entire image based on a first encryption key, and encrypts the same target area based on a composite key of the first encryption key and a second encryption key different from the first encryption key.

2. The sensor device according to claim 1, in, The detection unit performs a process of identifying a specific portion of the object, and The encryption unit performs encryption based on different encryption keys for an area of ​​the specific portion and other areas in the target area.

3. The sensor device according to claim 1, further comprising: an analyzing unit, analyzing the attribute or action of the target, and An output unit outputs information indicating a result of analysis performed by the analysis unit.

4. The sensor device according to claim 1, wherein: The detection unit performs detection of the target by decrypting the image signal encrypted based on the first encryption key using a just-in-time method.

5. The sensor device according to claim 1, wherein: The encryption unit encrypts the image signal based on a synthetic key obtained by synthesizing a plurality of encryption keys.

6. The sensor device according to claim 1, wherein: The encryption unit performs encryption for the target area while tracking the target.

7. The sensor device according to claim 1, wherein: The encryption unit encrypts the image signal using a stream cipher.

8. The sensor device according to claim 1, wherein: The array sensor, the detection unit, and the encryption unit are configured in one package.

9. The sensor device according to claim 1, wherein: The encryption unit generates an encryption key based on a photoelectric random number that is a random number obtained based on photoelectric conversion performed by the array sensor, and performs encryption of the image signal based on the generated encryption key.

10. The sensor device according to claim 1, wherein: The encryption unit encrypts read signals from pixels of the array sensor.

11. An encryption method comprising: detecting, based on an image signal obtained by imaging with an array sensor in which a plurality of pixels each having a light receiving element for visible light or invisible light are arranged one-dimensionally or two-dimensionally, a region of a target projected in an image as the target region; encrypting the entire image based on a first encryption key; as well as Encryption of the same target area in the image signal is performed based on information of the detected target area and based on a synthetic key of the first encryption key and a second encryption key different from the first encryption key.

Images