A method, device and readable storage medium for generating an image pulse sequence

Abstract

The application discloses a method and device for generating an image pulse sequence and a computer readable storage medium. The method comprises the following steps: acquiring the RGB channel grayscale value of each pixel in an image; calculating the hue value, saturation and lightness value of each pixel according to the RGB channel grayscale value; taking the hue value, saturation and lightness value of the current pixel as the parameters of a Poisson process respectively, taking the generated Poisson flow as the pulse sequence of the current pixel; and obtaining a pulse neural network model more in line with the working mechanism of the human brain, improving the recognition accuracy of the pulse neural network, and thus improving the light sensitivity.

Description

Technical Field

[0001] This application relates to the field of neural network model technology, specifically to a method, apparatus, and computer-readable storage medium for generating image pulse sequences. Background Technology

[0002] With the continuous development of computer technology, an Artificial Neural Network (ANN) algorithm has emerged that simulates the human brain's thought process. Spiking Neural Networks (SNNs) are a new generation of artificial neural network models inspired by biology. They belong to a subset of deep learning and have strong biological foundations. Building upon ANNs, SNNs use spiking neurons, which simulate the working mechanism of neurons in the brain, as activation units, mimicking the transmission process of biological neural information, giving SNNs a high degree of biological similarity. Unlike traditional artificial neural networks, spiking neurons use pulse sequences to represent information and typically possess memory functions. Spiking neurons have a more complex dynamic model, which allows SNNs to better utilize spatiotemporal information compared to ANNs, thus possessing stronger learning capabilities.

[0003] However, the current main way computers store and process images is to decompose each pixel in the image into a combination of RGB three colors. The pulse sequence generated based on the RGB channels has no biological basis and does not conform to the working mechanism of the human brain. Furthermore, the images in the dataset used for training SNN networks are of high quality, with high color contrast and little brightness fluctuation. In real-world scenarios with poor lighting, the output of the pulse sequence in each RGB channel will become sparse, and the spiking neurons will become difficult to activate. Summary of the Invention

[0004] To address the aforementioned technical problems, this application is proposed. Embodiments of this application provide a method, apparatus, and computer-readable storage medium for generating image pulse sequences, thereby resolving the aforementioned technical problems.

[0005] According to one aspect of this application, a method for generating an image pulse sequence is provided, applied to a spiking neural network model. The method for generating the image pulse sequence includes: obtaining the RGB channel grayscale value of each pixel in the image; calculating the hue value, saturation value, and brightness value of each pixel based on the RGB channel grayscale value; and using the hue value, saturation value, and brightness value of the current pixel as parameters of a Poisson process to generate a Poisson flow as the pulse sequence of the current pixel.

[0006] In one embodiment, the step of using the hue value, saturation, and brightness value of the current pixel as parameters of a Poisson process to generate a Poisson flow as a pulse sequence of the current pixel includes: normalizing the hue value of the current pixel to obtain a normalized hue value; and using the normalized hue value, saturation, and brightness value as parameters of a Poisson process to generate a Poisson flow as three pulse sequences of the current pixel.

[0007] In one embodiment, the normalization process for the hue value of the current pixel to obtain a normalized hue value includes: performing global normalization on the hue value of the current pixel, and / or performing sub-interval normalization on the hue value of the current pixel to obtain the normalized hue value.

[0008] In one embodiment, the global normalization process for the hue value of the current pixel includes: the normalized hue value Among them, H ij The hue value is mentioned.

[0009] In one embodiment, the sub-interval normalization process for the hue value of the current pixel includes: the normalized hue value Among them, H ij Let N be the hue value, N be the number of sub-intervals on the hue wheel, and m be the sub-interval number where the hue value is located.

[0010] In one embodiment, the step of using the normalized hue value, the saturation, and the brightness value as parameters of the Poisson process to generate a Poisson flow as three pulse sequences for the current pixel includes: using the normalized hue value as the probability value of the Poisson process to generate a Poisson flow as one pulse sequence for the current pixel.

[0011] In one embodiment, the step of using the normalized hue value, the saturation, and the brightness value as parameters of the Poisson process to generate a Poisson flow as three pulse sequences for the current pixel includes: using the saturation as the probability value of the Poisson process to generate a Poisson flow as one pulse sequence for the current pixel.

[0012] In one embodiment, the step of using the normalized hue value, saturation, and brightness value as parameters of a Poisson process to generate a Poisson flow as three pulse sequences for the current pixel includes: using the brightness value as the probability value of the Poisson process to generate a Poisson flow as one pulse sequence for the current pixel.

[0013] According to another aspect of this application, an image pulse sequence generation apparatus is provided, configured in a spiking neural network model. The image pulse sequence generation apparatus includes: a grayscale acquisition module for acquiring the RGB channel grayscale values ​​of each pixel in the image; a numerical conversion module for calculating the hue, saturation, and brightness values ​​of each pixel based on the RGB channel grayscale values; and a pulse generation module for using the hue, saturation, and brightness values ​​of the current pixel as parameters of a Poisson process to generate a Poisson flow as the pulse sequence of the current pixel.

[0014] According to another aspect of this application, a computer-readable storage medium is provided, characterized in that the storage medium stores a computer program for performing the image pulse sequence generation method described above.

[0015] This application provides a method, apparatus, and computer-readable storage medium for generating image pulse sequences. The method involves acquiring the RGB channel grayscale values ​​of each pixel in an image; calculating the hue, saturation, and brightness values ​​of each pixel based on the RGB channel grayscale values; and using the hue, saturation, and brightness values ​​of the current pixel as parameters for a Poisson process, generating a Poisson flow as the pulse sequence for the current pixel. Essentially, the method converts the RGB values ​​of each pixel in the image to HSV values, obtains a Poisson flow based on the Poisson process, and uses the Poisson flow as the pulse sequence required by a spiking neural network. This results in a spiking neural network model that better conforms to the working mechanism of the human brain, while simultaneously improving the recognition accuracy of the spiking neural network and thus enhancing light sensitivity. Attached Figure Description

[0016] The above and other objects, features, and advantages of this application will become more apparent from the more detailed description of the embodiments of this application in conjunction with the accompanying drawings. The drawings are provided to further illustrate the embodiments of this application and form part of the specification. They are used together with the embodiments of this application to explain this application and do not constitute a limitation thereof. In the drawings, the same reference numerals generally represent the same components or steps.

[0017] Figure 1 This is a schematic flowchart of an exemplary embodiment of the present application for a method of generating an image pulse sequence.

[0018] Figure 2 This is a flowchart illustrating a method for generating an image pulse sequence, provided in another exemplary embodiment of this application.

[0019] Figure 3 This is a schematic diagram illustrating the principle of the image pulse sequence generation method provided in this application.

[0020] Figure 4This is a schematic diagram of the structure of the pulse sequence during the execution of the image pulse sequence generation method provided in this application.

[0021] Figure 5 A schematic diagram illustrating the principle of HSV value calculation during the execution of the image pulse sequence generation method provided in this application.

[0022] Figure 6 This is a schematic diagram illustrating the principle of generating pulse sequences during the execution of the image pulse sequence generation method provided in this application.

[0023] Figure 7 This is a schematic diagram illustrating the principle of the pulse sequence input spiking neural network model during the execution of the image pulse sequence generation method provided in this application.

[0024] Figure 8 This is a schematic diagram of the structure of an image pulse sequence generation apparatus provided in an exemplary embodiment of this application.

[0025] Figure 9 This is a schematic diagram of the structure of an image pulse sequence generation apparatus provided in another exemplary embodiment of this application.

[0026] Figure 10 This is a structural diagram of an electronic device provided in an exemplary embodiment of this application. Detailed Implementation

[0027] Hereinafter, exemplary embodiments according to this application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are merely some embodiments of this application, and not all embodiments of this application. It should be understood that this application is not limited to the exemplary embodiments described herein.

[0028] The primary storage and processing method for color images in computers is to decompose each pixel into a combination of the three primary colors: red, green, and blue (RGB). Each basic color unit has 256 color levels, and a single pixel can be represented by three eight-bit binary numbers. When the value of the current pixel is obtained, if it is converted into a binary number, the original binary number is directly output as an eight-bit binary sequence. If the value is processed as a real number, a stateful delay encoder or periodic encoder is used to output a pulse stream, or a stateless Poisson encoder is used to generate a pulse sequence from the original data. However, the pulse sequence generated based on RGB channel encoding focuses more on the grayscale features of the image, which is not how the human brain works. Furthermore, SNN networks typically select high-quality images as training samples during training, which have high color contrast and low brightness fluctuations. In real-world scenarios, poor lighting can cause the RGB channel pulse sequence to become sparse, making it difficult to activate the spiking neurons of the SNN network.

[0029] To address the aforementioned technical problems, this application proposes a method and apparatus for generating image pulse sequences. By converting the RGB value of each pixel in the image into HSV (Hue, Saturation, Value) values, specifically including hue (H), saturation (S), and brightness (V), and obtaining the pulse sequence of the SNN network based on Bernoulli experiments, the recognition accuracy of the spiking neural network is improved, thereby enhancing light sensitivity and making the spiking neural network more consistent with the working mechanism of the human brain when processing image data.

[0030] The specific structure and implementation of the image pulse sequence generation method and apparatus provided in this application will be described in detail below with reference to the accompanying drawings.

[0031] Figure 1 This is a schematic flowchart illustrating a method for generating an image pulse sequence according to an exemplary embodiment of this application. This method for generating an image pulse sequence is applied to a spiking neural network model, such as... Figure 1 As shown, the method for generating this image pulse sequence includes the following steps:

[0032] Step 110: Obtain the grayscale value of the RGB channel for each pixel in the image.

[0033] By acquiring an image and obtaining the RGB value of each pixel in the image, we can obtain the characteristic information of each pixel.

[0034] Step 120: Calculate the hue, saturation, and brightness values ​​of each pixel based on the grayscale values ​​of the RGB channels.

[0035] After determining the grayscale values ​​of the RGB channels of a pixel, the hue, saturation, and brightness values ​​of that pixel are calculated based on the grayscale values ​​of each channel.

[0036] First, calculate the brightness value of the current pixel (the pixel in the i-th row and j-th column). Specifically, the brightness value can be calculated using the following formula:

[0037] V ij =max(R) ij G ij B ij ) / 255;

[0038] Among them, R ij G ij B ij These are the grayscale values ​​for the three RGB channels, respectively.

[0039] Then, calculate the saturation of the current pixel (the pixel in the i-th row and j-th column). Specifically, first determine the grayscale range of the current pixel's channels:

[0040] Δij =(max(R) ij G ij B ij )-min(R ij G ij B ij )) / 255;

[0041] The saturation of the current pixel can be calculated using the following formula:

[0042] S ij =Δ ij / V ij In particular, when V ij When S is 0, ij =0.

[0043] Finally, calculate the hue value of the current pixel (the pixel in the i-th row and j-th column). Specifically, the hue value can be calculated using the following formula:

[0044] In particular, when Δ ij When H is 0, ij =0.

[0045] Step 130: Use the hue, saturation, and brightness values ​​of the current pixel as parameters for the Poisson process to generate a Poisson flow as the pulse sequence of the current pixel.

[0046] After calculating the hue, saturation, and brightness values ​​of the current pixel, the hue, saturation, and brightness values ​​are used as parameters of the Poisson process (through a normal encoder) to generate a Poisson flow as a pulse sequence, thereby obtaining a pulse sequence that is more in line with the way the human brain recognizes images, thus improving the image recognition accuracy and sensitivity of the spiking neural network model.

[0047] This application provides a method for generating an image pulse sequence. The method involves acquiring the RGB channel grayscale values ​​of each pixel in an image; calculating the hue, saturation, and brightness values ​​of each pixel based on the RGB channel grayscale values; and using the hue, saturation, and brightness values ​​of the current pixel as parameters for a Poisson process to generate a Poisson flow as the pulse sequence of the current pixel. Essentially, the method converts the RGB values ​​of each pixel in the image to HSV values ​​and generates a Poisson flow based on the Poisson process to obtain the pulse sequence required by a spiking neural network. This results in a spiking neural network model that better conforms to the working mechanism of the human brain, while simultaneously improving the recognition accuracy of the spiking neural network and thus enhancing light sensitivity.

[0048] Figure 2 This is a schematic flowchart illustrating a method for generating an image pulse sequence according to another exemplary embodiment of this application. Figure 2 As shown, step 130 above may include:

[0049] Step 131: Normalize the hue value of the current pixel to obtain the normalized hue value.

[0050] The hue value calculated by the above method is usually greater than 1. In order to generate a pulse sequence based on the hue value, the hue value can be normalized to convert the hue value into a value between 0 and 1. Thus, the hue value, saturation and brightness value between 0 and 1 can be obtained. Then, it can be encoded by a normal encoder to obtain a pulse sequence suitable for a spiking neural network model.

[0051] In one embodiment, step 131 may be implemented by performing global normalization on the hue value of the current pixel, and / or performing sub-interval normalization on the hue value of the current pixel to obtain the normalized hue value.

[0052] Optionally, the specific implementation of global normalization of the hue value of the current pixel can be: the normalized hue value Among them, H ij This is the hue value.

[0053] The maximum degree of the tone wheel is 360 degrees. Dividing the original tone value of the current pixel by 360 normalizes the original tone value of the current pixel to a value between 0 and 1, so that a normal encoder can be used to encode it to obtain a pulse sequence suitable for a spiking neural network model.

[0054] Optionally, the specific implementation of sub-interval normalization of the hue value of the current pixel can be: the normalized hue value Among them, H ij is the hue value, N is the number of sub-intervals on the hue wheel, and m is the sub-interval number where the hue value is located.

[0055] The tone wheel can be evenly divided into multiple sub-intervals of varying numbers. Based on the requirements of the actual task, the original tone value of the current pixel is normalized within the corresponding sub-interval to normalize the original tone value of the current pixel to a value between 0 and 1. This allows for encoding using a normal encoder to obtain a pulse sequence suitable for a spiking neural network model. Preferably, when using sub-interval normalization, the tone position information α is calculated. The value of α is determined by the following formula: α = m / N, and based on the normalized tone value... And α generate a pulse sequence.

[0056] Step 132: Use the normalized hue, saturation, and brightness values ​​as parameters for the Poisson process to generate a Poisson flow as the three pulse sequences for the current pixel.

[0057] After normalizing the hue value, the hue value, saturation value, and brightness value are all values ​​between 0 and 1. At this point, the hue value, saturation value, and brightness value can be used as parameters of the Poisson process to generate an n-dimensional pulse sequence suitable for a spiking neural network model.

[0058] In one embodiment, step 132 can be implemented by using the normalized hue value as the probability value of the Poisson process (the probability value of the pulse output being 1) to obtain an n-dimensional pulse sequence.

[0059] Specifically, after calculating and normalizing the hue value, the normalized hue value (between 0 and 1) is used as the probability value of the Poisson process (the probability value of the pulse output being 1) to obtain an n-dimensional pulse sequence.

[0060] In one embodiment, step 132 can be implemented by using saturation as the probability value of the Poisson process (the probability value of the pulse output being 1) to obtain an n-dimensional pulse sequence.

[0061] Specifically, after calculating the saturation, the value of the saturation (between 0 and 1) is used as the probability value of the Poisson process (the probability value of the pulse output being 1) to obtain an n-dimensional pulse sequence.

[0062] In one embodiment, step 132 can be implemented by performing n Poisson processes using the brightness value as the probability value of the pulse output being 1, to obtain an n-dimensional pulse sequence.

[0063] Specifically, after calculating the brightness value, the brightness value (between 0 and 1) is used as the probability value for the pulse output to be 1 to perform n Poisson processes to obtain an n-dimensional pulse sequence.

[0064] Figure 3 This is a schematic diagram illustrating the principle of the image pulse sequence generation method provided in this application. Figure 4 This is a schematic diagram illustrating the structure of the pulse sequence during the execution of the image pulse sequence generation method provided in this application. (See attached diagram.) Figure 3 and Figure 4 As shown, the HSV value is calculated based on the RGB values ​​of the pixels in the image, and a pulse sequence of 0 or 1 is generated based on the HSV value. The HSV value is positively correlated with the probability of 1 appearing in the pulse sequence (when the HSV value is between 0 and 1, the probability of 1 appearing in the pulse sequence is equal to the HSV value).

[0065] The following example illustrates the specific process of the image pulse sequence production method provided in this application. Figure 5 This is a schematic diagram illustrating the principle of HSV value calculation during the execution of the image pulse sequence generation method provided in this application. Figure 6This is a schematic diagram illustrating the principle of generating pulse sequences during the execution of the image pulse sequence generation method provided in this application. Figure 7 This is a schematic diagram illustrating the principle of the pulse sequence input spiking neural network model during the execution of the image pulse sequence generation method provided in this application. Figure 5 As shown, taking a 2x2 image as an example, the following explanation only uses the first pixel as the current pixel. The RGB values ​​of the current pixel are: R=100, G=150, B=200. First, calculate the brightness value V=200 / 255=0.78, and then calculate the channel grayscale range Δ. ij = (200-100) / 255 = 0.39, and calculate the saturation S = 0.39 / 0.78 = 0.5 based on the channel grayscale range. Finally, calculate the hue value H = 60*(100-150) / 100+240 = 210°, and after global normalization of the hue value, we get H* = 210 / 360 = 0.58. After sub-interval normalization of the hue value, we get H1 = 1-(2-3*210 / 360) = 0.75, and the hue position H2 = 1 / 3.

[0066] like Figure 6 As shown, after calculating the HSV value, the HSV value is used as the probability value of the Poisson process (the probability of the output being 1) and the process is repeated n times (e.g., Figure 6 As shown, to obtain a sequence length of 32, 32 Poisson processes are performed to obtain an n-dimensional pulse sequence. It should be understood that this application can generate pulse sequences using only globally normalized hue values, or using sub-interval normalized hue values, or simultaneously using both globally normalized and sub-interval normalized hue values. Figure 7 As shown, after obtaining the pulse sequence, the pulse sequence can be input into the spiking neural network model to train the spiking neural network model, thereby improving the recognition accuracy of the spiking neural network model for images.

[0067] Figure 8 This is a schematic flowchart illustrating a method for generating an image pulse sequence according to an exemplary embodiment of this application. The device for generating the image pulse sequence is configured within a spiking neural network model, such as... Figure 8 As shown, the image pulse sequence generation device 80 includes: a grayscale acquisition module 81, used to acquire the RGB channel grayscale value of each pixel in the image; a numerical conversion module 82, used to calculate the hue value, saturation value and brightness value of each pixel based on the RGB channel grayscale value; and a pulse generation module 83, used to use the hue value, saturation value and brightness value of the current pixel as parameters of the Poisson process to generate a Poisson flow as the pulse sequence of the current pixel.

[0068] This application provides an image pulse sequence generation device. A grayscale acquisition module 81 acquires the RGB channel grayscale value of each pixel in the image; a numerical conversion module 82 calculates the hue, saturation, and brightness values ​​of each pixel based on the RGB channel grayscale values; and a pulse generation module 83 uses the hue, saturation, and brightness values ​​of the current pixel as parameters for a Poisson process to generate a Poisson flow as the pulse sequence of the current pixel. In essence, the RGB values ​​of each pixel in the image are converted to HSV values, and the pulse sequence required by the spiking neural network is obtained based on the Poisson process. This results in a spiking neural network model that better conforms to the working mechanism of the human brain, while simultaneously improving the recognition accuracy of the spiking neural network and thus enhancing light sensitivity.

[0069] Figure 9 This is a schematic flowchart illustrating a method for generating an image pulse sequence according to another exemplary embodiment of this application. Figure 9 As shown, the pulse generation module 83 may include: a normalization unit 831, used to normalize the hue value of the current pixel to obtain a normalized hue value; and a sequence generation unit 832, used to use the normalized hue value, saturation and brightness value as parameters of the Poisson process to generate a Poisson flow as three n-dimensional pulse sequences of the current pixel.

[0070] In one embodiment, the normalization unit 831 may be further configured to: perform global normalization processing on the hue value of the current pixel, and / or perform sub-interval normalization processing on the hue value of the current pixel, so as to obtain the normalized hue value.

[0071] In one embodiment, the sequence generation unit 832 may be further configured to: use the normalized hue value as the probability value of the Poisson process (the probability value of the pulse output being 1), perform n Poisson processes to obtain an n-dimensional pulse sequence.

[0072] In one embodiment, the sequence generation unit 832 may be further configured to perform n Poisson processes using saturation as the probability value of the Poisson process (the probability value of the pulse output being 1) to obtain an n-dimensional pulse sequence.

[0073] In one embodiment, the sequence generation unit 832 may be further configured to perform n Poisson processes using the brightness value as the probability value of the Poisson process (the probability value of the pulse output being 1) to obtain an n-dimensional pulse sequence.

[0074] Below, for reference Figure 10 This application describes an electronic device according to embodiments thereof. The electronic device may be either or both of a first device and a second device, or a standalone device independent of them, which may communicate with the first device and the second device to receive acquired input signals from them.

[0075] Figure 10 A block diagram of an electronic device according to an embodiment of this application is illustrated.

[0076] like Figure 10 As shown, the electronic device 10 includes one or more processors 11 and memory 12.

[0077] The processor 11 may be a central processing unit (CPU) or other form of processing unit with data processing capabilities and / or instruction execution capabilities, and may control other components in the electronic device 10 to perform desired functions.

[0078] The memory 12 may include one or more computer program products, which may include various forms of computer-readable storage media, such as volatile memory and / or non-volatile memory. The volatile memory may include, for example, random access memory (RAM) and / or cache memory. The non-volatile memory may include, for example, read-only memory (ROM), hard disk, flash memory, etc. One or more computer program instructions may be stored on the computer-readable storage medium, and the processor 11 may execute the program instructions to implement the methods of the various embodiments of this application described above and / or other desired functions. Various contents such as input signals, signal components, and noise components may also be stored in the computer-readable storage medium.

[0079] In one example, the electronic device 10 may also include an input device 13 and an output device 14, which are interconnected via a bus system and / or other forms of connection mechanism (not shown).

[0080] When the electronic device is a standalone device, the input device 13 can be a communication network connector for receiving the collected input signals from the first device and the second device.

[0081] In addition, the input device 13 may also include, for example, a keyboard, a mouse, etc.

[0082] The output device 14 can output various information to the outside, including determined distance information, direction information, etc. The output device 14 may include, for example, a display, a speaker, a printer, and a communication network and its connected remote output devices, etc.

[0083] Of course, for the sake of simplicity, Figure 10 Only some of the components of the electronic device 10 relevant to this application are shown in this illustration; components such as buses, input / output interfaces, etc., are omitted. In addition, the electronic device 10 may include any other suitable components depending on the specific application.

[0084] The computer program product can be written in any combination of one or more programming languages ​​to perform the operations of the embodiments of this application. The programming languages ​​include object-oriented programming languages ​​such as Java and C++, as well as conventional procedural programming languages ​​such as C or similar languages. The program code can be executed entirely on the user's computing device, partially on the user's computing device, as a standalone software package, partially on the user's computing device and partially on a remote computing device, or entirely on a remote computing device or server.

[0085] The computer-readable storage medium may be any combination of one or more readable media. A readable medium may be a readable signal medium or a readable storage medium. A readable storage medium may, for example, include, but is not limited to, electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatuses, or devices, or any combination thereof. More specific examples of readable storage media (a non-exhaustive list) include: electrical connections having one or more wires, portable disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.

[0086] The above description has been given for purposes of illustration and description. Furthermore, this description is not intended to limit the embodiments of this application to the forms disclosed herein. Although numerous exemplary aspects and embodiments have been discussed above, those skilled in the art will recognize certain variations, modifications, alterations, additions, and sub-combinations thereof.

Claims

1. A method for generating an image pulse sequence, characterized in that, Applied to a spiking neural network model, the method for generating the image pulse sequence includes: Obtain the grayscale values of the RGB channels of each pixel in the image; Calculate the hue value, saturation, and lightness value of each pixel according to the grayscale values of the RGB channels; and Respectively use the hue value, the saturation, and the lightness value of the current pixel as the parameters of a Poisson process to generate a Poisson flow as the pulse sequence of the current pixel; where The step of respectively using the hue value, the saturation, and the lightness value of the current pixel as the parameters of a Poisson process to generate a Poisson flow as the pulse sequence of the current pixel includes: Normalize the hue value of the current pixel to obtain a normalized hue value; and Respectively use the normalized hue value, the saturation, and the lightness value as the parameters of a Poisson process to generate a Poisson flow as three pulse sequences of the current pixel.

2. The method for generating an image pulse sequence according to claim 1, characterized in that, The step of normalizing the hue value of the current pixel to obtain a normalized hue value includes: Perform global normalization on the hue value of the current pixel, and / or perform sub-interval normalization on the hue value of the current pixel to obtain the normalized hue value.

3. The method for generating an image pulse sequence according to claim 2, characterized in that, The step of performing global normalization on the hue value of the current pixel includes: The normalized hue value ; in, The hue value is mentioned.

4. The method for generating an image pulse sequence according to claim 2, characterized in that, The step of performing sub-interval normalization on the hue value of the current pixel includes: The normalized hue value ; in, For the hue value, This represents the number of sub-ranges on the color wheel. The sub-interval number where the hue value is located.

5. The method for generating an image pulse sequence according to claim 1, characterized in that, The step of respectively using the normalized hue value, the saturation, and the lightness value as the parameters of a Poisson process to generate a Poisson flow as three pulse sequences of the current pixel includes: Use the normalized hue value as the probability value of the Poisson process to generate a Poisson flow as one pulse sequence of the current pixel.

6. The method for generating an image pulse sequence according to claim 1, characterized in that, The step of respectively using the normalized hue value, the saturation, and the lightness value as the parameters of a Poisson process to generate a Poisson flow as three pulse sequences of the current pixel includes: Use the saturation as the probability value of the Poisson process to generate a Poisson flow as one pulse sequence of the current pixel.

7. The method for generating an image pulse sequence according to claim 1, characterized in that, The step of respectively using the normalized hue value, the saturation, and the lightness value as the parameters of a Poisson process to generate a Poisson flow as three pulse sequences of the current pixel includes: Use the lightness value as the probability value of the Poisson process to generate a Poisson flow as one pulse sequence of the current pixel.

8. An apparatus for generating an image pulse sequence, characterized in that, Applied to a spiking neural network model, the apparatus for generating the image pulse sequence includes: A grayscale acquisition module, configured to obtain the grayscale values of the RGB channels of each pixel in the image; A numerical conversion module, configured to calculate the hue value, saturation, and lightness value of each pixel according to the grayscale values of the RGB channels; and A pulse generation module is used to take the hue value, saturation value, and brightness value of the current pixel as parameters of a Poisson process to generate a Poisson flow as a pulse sequence for the current pixel; wherein, the pulse generation module includes a normalization unit, used to normalize the hue value of the current pixel to obtain a normalized hue value, and to take the normalized hue value, saturation value, and brightness value as parameters of a Poisson process to generate a Poisson flow as three pulse sequences for the current pixel.

9. A computer-readable storage medium, characterized in that, The storage medium stores a computer program for executing the image pulse sequence generation method according to any one of claims 1-7.

Images