Image classification method, system and terminal based on DCPI resin

Abstract

The invention discloses an image classification method, system and terminal based on DCPI resin, and the method comprises the steps: constructing a photon neural network based on DCPI resin, and the photon neural network comprises an input layer, an optical weight self-learning layer and an output layer; training the photon neural network to obtain a target network; collecting a target image, inputting the target image into an input layer of the target network for preprocessing to obtain a preprocessed image, inputting the preprocessed image into an optical weight self-learning layer of the target network for reasoning to obtain a reasoning result, and inputting the reasoning result into an output layer of the target network for classification to obtain a classification result. And outputting a classification result of the target image. The photon neural network based on the DCPI resin is an independent photon neural network which does not depend on an artificial neural network, the generation of the weight does not need a large amount of computing resources and energy consumption, and the method is an important step of human-like artificial intelligence.

Description

Technical Field

[0001] The present invention relates to the technical field of photon neural network, and in particular to an image classification method, system, terminal and computer-readable storage medium based on DCPI resin. Background Art

[0002] With the booming development of integrated circuit technology, artificial intelligence has rapidly become one of the most active fields in modern society. However, due to the slowdown of Moore's Law, the von Neumann bottleneck of existing computer architectures, and inherent defects such as processing speed being proportional to energy loss, people have once again turned their attention to the field of optics. They hope to build neural networks in the field of optics, and use the parallel processing capabilities and high speed and low power consumption of light to overcome the various limitations faced by electronic neural networks, thereby promoting new breakthroughs in artificial intelligence.

[0003] Photonic neural network is an artificial neural network that uses optical technology to process information. It performs calculations through photons instead of electrons, has the ability to process and transmit information in super parallel, and is suitable for scenarios with high-density leads and direct image processing. The development of photonic neural network technology has largely drawn on artificial neural networks.

[0004] However, when it comes to the critical issue of obtaining weights, photonic neural networks still rely on artificial neural networks and do not have their own methods. Moreover, the method by which ANN obtains weights is based on the back-propagation algorithm, which is not only extremely computationally intensive, but also energy intensive, requiring a large amount of computing resources and energy consumption to achieve its complex calculation and processing requirements. This results in high energy consumption in the image classification process currently achieved through photonic neural networks, which runs counter to the expectation of ONN to achieve "low power consumption".

[0005] Therefore, the prior art still needs to be improved and developed. Summary of the invention

[0006] The main purpose of the present invention is to provide an image classification method, system, terminal and computer-readable storage medium based on DCPI resin, aiming to solve the problem that the photonic neural network in the prior art lacks a technology that can obtain weights with low power consumption, resulting in high energy consumption in the image classification process currently implemented by the photonic neural network.

[0007] To achieve the above object, the present invention provides an image classification method based on DCPI resin, and the image classification method based on DCPI resin comprises the following steps:

[0008] Constructing a photonic neural network based on DCPI resin, wherein the photonic neural network includes an input layer, an optical weight self-learning layer and an output layer;

[0009] Training the photonic neural network to obtain a target network;

[0010] Capture a target image, input the target image into the input layer of the target network for preprocessing to obtain a preprocessed image, input the preprocessed image into the optical weight self-learning layer of the target network for inference to obtain an inference result, input the inference result into the output layer of the target network for classification, and output the classification result of the target image.

[0011] Optionally, the DCPI resin-based image classification method, wherein the optical weight self-learning layer includes a UV projection module, a first microlens array module, a DCPI resin module, a second microlens array module and a laser irradiation module;

[0012] Wherein, the DCPI resin module is located between the first microlens array module and the second microlens array module.

[0013] Optionally, in the DCPI resin-based image classification method, the DCPI resin module is obtained by imitating a biological learning process.

[0014] Optionally, the DCPI resin-based image classification method, wherein the photonic neural network is trained to obtain a target network, specifically comprises:

[0015] Acquire a sample image, and preprocess the sample image to obtain a preprocessed sample image;

[0016] The preprocessed sample image is input into the UV projection module of the optical weight self-learning layer, and the preprocessed sample image is irradiated onto the DCPI resin module after passing through the first microlens array module by the first wavelength laser emitted by the UV projection module to obtain a first stimulation signal;

[0017] The second wavelength laser emitted by the laser irradiation module passes through the second microlens array module and then irradiates the DCPI resin module to obtain a second stimulation signal;

[0018] The first stimulation signal and the second stimulation signal are associated with each other through the DCPI resin module, and the weight of the information carried on the image is automatically generated, and the target network is obtained according to the weight.

[0019] Optionally, the DCPI resin-based image classification method, wherein the step of inputting the target image into the input layer of the target network for preprocessing to obtain a preprocessed image, specifically comprises:

[0020] Inputting the target image into the input layer of the target network for denoising, filtering and normalization to obtain first intermediate data;

[0021] Compressing the first intermediate data and removing irrelevant information to reduce the number of weights to be learned by the optical weight self-learning layer, thereby obtaining second intermediate data;

[0022] The size of the second intermediate data is scaled according to the size of the optical weight self-learning layer to obtain third intermediate data, and initial feature extraction is performed on the third intermediate data to obtain the preprocessed image.

[0023] Optionally, the DCPI resin-based image classification method, wherein the preprocessed image is input into the optical weight self-learning layer of the target network for reasoning to obtain the reasoning result, specifically includes:

[0024] Inputting the preprocessed image into a weighted optical weight self-learning layer, super-projecting the preprocessed image through the rays emitted by the UV projection module, and obtaining a green fluorescence signal;

[0025] The fluorescent signal is converted into an electrical signal, and an inference result is obtained according to the electrical signal.

[0026] Optionally, in the DCPI resin-based image classification method, the first wavelength laser emitted by the UV projection module is a violet laser with a wavelength of 365 nm, and the second wavelength laser emitted by the laser irradiation module is a green laser with a wavelength of 532 nm.

[0027] In addition, to achieve the above-mentioned purpose, the present invention also provides an image classification method system based on DCPI resin, wherein the image classification method system based on DCPI resin comprises:

[0028] A model building module, used to build a photonic neural network based on DCPI resin, wherein the photonic neural network includes an input layer, an optical weight self-learning layer and an output layer;

[0029] A model training module, used to train the photonic neural network to obtain a target network;

[0030] The optical computing module is used to capture a target image, input the target image into the input layer of the target network for preprocessing to obtain a preprocessed image, input the preprocessed image into the optical weight self-learning layer of the target network for inference to obtain an inference result, input the inference result into the output layer of the target network for classification, and output the classification result of the target image.

[0031] In addition, to achieve the above-mentioned purpose, the present invention also provides a terminal, wherein the terminal includes: a memory, a processor, and a DCPI resin-based image classification method program stored in the memory and executable on the processor, wherein the DCPI resin-based image classification method program implements the steps of the DCPI resin-based image classification method as described above when executed by the processor.

[0032] In addition, to achieve the above-mentioned purpose, the present invention also provides a computer-readable storage medium, wherein the computer-readable storage medium stores a DCPI resin-based image classification method program, and when the DCPI resin-based image classification method program is executed by a processor, the steps of the DCPI resin-based image classification method as described above are implemented.

[0033] In the present invention, a photon neural network based on DCPI resin is constructed, and the photon neural network includes an input layer, an optical weight self-learning layer and an output layer; the photon neural network is trained to obtain a target network; a target image is collected, and the target image is input into the input layer of the target network for preprocessing to obtain a preprocessed image, and the preprocessed image is input into the optical weight self-learning layer of the target network for reasoning to obtain a reasoning result, and the reasoning result is input into the output layer of the target network for classification, and the classification result of the target image is output. The photon neural network based on DCPI resin in the present invention is an independent photon neural network that does not rely on artificial neural networks. The generation of weights does not require a large amount of computing resources and energy consumption, and is an important step in human-like artificial intelligence. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Figure 1 It is a flow chart of a preferred embodiment of the image classification method based on DCPI resin of the present invention;

[0035] Figure 2 It is a schematic diagram of the optical weight generation and reasoning process in the image classification method based on DCPI resin of the present invention;

[0036] Figure 3 is another schematic diagram of the optical weight generation and reasoning process in the DCPI resin-based image classification method of the present invention;

[0037] Figure 4 Schematic diagram of using light of different wavelengths to irradiate DCPI resin in the DCPI resin-based image classification method of the present invention;

[0038] Figure 5 It is a structural diagram of a preferred embodiment of the image classification method system based on DCPI resin of the present invention;

[0039] Figure 6Schematic diagram of the operating environment of a preferred embodiment of the terminal of the present invention. DETAILED DESCRIPTION

[0040] The present application provides an image classification method, system and terminal based on DCPI resin. In order to make the purpose, technical solution and effect of the present application clearer and more specific, the present application is further described in detail with reference to the accompanying drawings and examples. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0041] It will be understood by those skilled in the art that, unless otherwise defined, all terms (including technical and scientific terms) used herein have the same meaning as those generally understood by those skilled in the art to which this application belongs. It should also be understood that terms such as those defined in general dictionaries should be understood to have meanings consistent with those in the context of the prior art, and will not be interpreted with idealized or overly formal meanings unless specifically defined as here.

[0042] In addition, if there are descriptions involving "first", "second", etc. in the embodiments of the present invention, the descriptions of "first", "second", etc. are only used for descriptive purposes and cannot be understood as indicating or suggesting their relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include at least one of the features. In addition, the technical solutions between the various embodiments can be combined with each other, but they must be based on the ability of ordinary technicians in the field to implement them. When the combination of technical solutions is contradictory or cannot be implemented, it should be deemed that such a combination of technical solutions does not exist and is not within the scope of protection required by the present invention.

[0043] The image classification method based on DCPI resin described in the preferred embodiment of the present invention is as follows: Figure 1 As shown, the image classification method based on DCPI resin comprises the following steps:

[0044] Step S10, constructing a photonic neural network based on DCPI resin, wherein the photonic neural network includes an input layer, an optical weight self-learning layer and an output layer.

[0045] It is understandable that a complete photonic neural network should have three modules: input, processing and output. In this embodiment, the input layer and the output layer have been built long ago, and the focus of the present invention is the processing module, that is, the optical weight self-learning layer.

[0046] like Figure 2As shown, the optical weight self-learning layer is used for weight self-learning and reasoning, and the optical weight self-learning layer includes a UV projection module, a first microlens array module, a DCPI resin module, a second microlens array module and a laser irradiation module; wherein the DCPI resin module is located between the first microlens array module and the second microlens array module.

[0047] Among them, the UV (Ultraviolet) projection module and the laser irradiation module are used to emit rays of different wavelengths, the first microlens array module and the second microlens array module are used to receive the emitted rays and focus and image them, and the DCPI resin (Dual Color Photo initiator) module is used to receive light from two different directions and associate the dual light stimulation with the fluorescence spectrum change of the resin, thereby realizing the automatic generation of weights.

[0048] Furthermore, the DCPI resin module is obtained by imitating a biological learning process; wherein, the imitating biological learning process is to use the dual-light fluorescence effect to analogize the Pavlov and dog experiment, implement the Hebb rule through an optical experiment, and associate the dual-light stimulation with the fluorescence spectrum change of the resin to achieve automatic generation of weights.

[0049] It is understandable that for the acquisition of weights, the present invention proposes a more radical method, which directly imitates the learning process of organisms, such as Pavlov's experiment with dogs. This is a study of classical conditioning, which shows how animals establish conditioned reflexes through learning. Knowing that dogs will intrinsically secrete saliva when they see food, Pavlov rang a bell before each feeding, and eventually made the dog secrete saliva when it heard the bell even if there was no food, which proved that animals can form new behavior patterns through repeated stimulus-response processes. This process involves the brain associating two originally independent stimuli with each other in the memory system, so that the reaction originally triggered by a specific stimulus can be triggered by another new, independent stimulus. These stimuli can enter the brain through different sensory channels, such as vision and hearing, and establish connections in the brain after complex neural processing. Finally, the brain can flexibly integrate information from different sources into a unified memory representation. The memory formed by the brain associating the two together is the weight in the physical sense of the present invention.

[0050] Step S20: training the photonic neural network to obtain a target network.

[0051] Specifically, a sample image is acquired, and the sample image is preprocessed to obtain a preprocessed sample image.

[0052] In this embodiment, the sample image is input into the input layer of the target network for denoising, filtering and normalization, and the processed data is compressed and irrelevant information is removed to reduce the number of weights to be learned by the optical weight self-learning layer. The size of the data after removing irrelevant information is scaled according to the size of the optical weight self-learning layer, and initial feature extraction is performed on the scaled data to obtain a preprocessed sample image.

[0053] like Figure 2 As shown, the preprocessed sample image is input into the UV projection module of the optical weight self-learning layer, and the preprocessed sample image is irradiated onto the DCPI resin module after passing through the first microlens array module by the first wavelength laser emitted by the UV projection module to obtain a first stimulation signal; the second wavelength laser emitted by the laser irradiation module is irradiated onto the DCPI resin module after passing through the second microlens array module to obtain a second stimulation signal.

[0054] In this embodiment, the key to the model lies in two independent stimulation signals (a first stimulation signal and a second stimulation signal) and a hardware (DCPI resin) capable of simultaneously associating two independent input signals. Specifically, the first wavelength laser emitted by the UV projection module irradiates the preprocessed sample image onto the DCPI resin module after passing through the first microlens array module, and the second wavelength laser emitted by the laser irradiation module passes through the second microlens array module and irradiates onto the DCPI resin module to obtain a second stimulation signal.

[0055] The first wavelength laser emitted by the UV projection module is a violet laser with a wavelength of 365 nm, and the second wavelength laser emitted by the laser irradiation module is a green laser with a wavelength of 532 nm.

[0056] Furthermore, the first stimulation signal and the second stimulation signal are associated with each other through the DCPI resin module, and a weight of information carried on the image is automatically generated, and the target network is obtained according to the weight.

[0057] It can be understood that the dual-light fluorescence effect based on DCPI resin shows excellent correlation characteristics. It is manifested in that when the DCPI resin is irradiated with a laser of 532 wavelength (green light), the stimulated reaction is seen by the human eye as emitting green light, and this stimulus is represented as an unconditional stimulus (the dog secretes saliva when seeing food). When the DCPI resin is irradiated with light of 365 wavelength, the stimulated reaction is seen by the human eye as emitting red light, and this stimulus is represented as a neutral stimulus (the dog will not secrete saliva when hearing the sound of a bell). When the DCPI resin is irradiated with light of 365 and 532 wavelengths at the same time and then irradiated with light of 365 wavelength, the stimulated reaction is seen by the human eye as emitting green light. At this time, the stimulation of the two lights (unconditional and neutral) has been linked (the dog will secrete saliva when hearing the sound of a bell). Therefore, DCPI resin can become an associative element. Using the dual-light fluorescence effect to analogize Pavlov and the dog experiment, the Hebb rule is successfully implemented through optical experiments, and the dual-light stimulation is associated with the change of the resin's fluorescence spectrum, thereby realizing the automatic generation of weights, and being able to complete the typical artificial intelligence task-pattern recognition.

[0058] Furthermore, if Figure 3 As shown, in another embodiment, the samples to be learned include the letters NVZ. At this time, a 365nm (purple) projector is used to project these three letters onto the DCPI. When projecting the letter V, only the upper area is irradiated with 532nm (green) light, while there is no 532nm light in the middle and below. At this time, because only the resin in the upper area is irradiated with 532nm and 365nm, the fluorescence spectrum changes. Therefore, when only 365nm is irradiated, it can be seen that only the red letter V appears in the upper area. Continue to use the 365nm projector to project the letter N. Only the middle area is irradiated with 532nm light, while there is no 532nm light in the upper and lower areas. At this time, because only the resin in the middle area is irradiated with 532nm and 365nm, the fluorescence spectrum changes. Therefore, when only 365nm is irradiated, it can be seen that only the red letter N appears in the middle area. Continue to use the 365nm projector to project the letter Z. Only the lower area is irradiated with 532nm light, while the upper and middle areas are not irradiated with 532nm light. At this time, because only the resin in the lower area is irradiated with 532nm and 365nm, the fluorescence spectrum changes. Therefore, when only 365nm is irradiated, you can see that only the red letter N appears in the lower area. After the above steps, the weights of three letters are engraved on the DCPI.

[0059] It is understandable that the traditional computer digital weights are 0 and 1. In the present invention, the weight is distinguished by the light that is stimulated, for example: 1 is stimulated by green light, and 0 is stimulated by red light. When DCPI is irradiated with only 365nm light, it stimulates red light 0. When 365+532 irradiates DCPI for a period of time, and then irradiates with 365 alone, DCPI stimulates green light 1. When 365 light is used to project letters, such as projecting the letter V, a red pattern of the letter V will appear in DCPI. At this time, a 532 light is used to cover the entire letter area. After 365+532 irradiation for a period of time, 365 light is used alone to irradiate, and a green pattern of the letter V appears. As a result, the light stimulated by DCPI is converted from red light 0 to green light 1. Through this method, the weight corresponding to the letter can be engraved on DCPI.

[0060] At this time, continue to use the 365nm projector to project the three letters N, V, and Z respectively. According to the DCPI resin after the fluorescence spectrum changes when irradiated with 365nm light alone, it is known that green fluorescence is stimulated. When the letter N is projected, the green fluorescence stimulated by the weight corresponding to the letter N is the largest. When the letter V is projected, the green fluorescence stimulated by the weight corresponding to the letter V is the largest. When the letter Z is projected, the green fluorescence stimulated by the weight corresponding to the letter Z is the largest. Therefore, the dual-light fluorescence effect is used to analogize Pavlov's experiment with the dog, and the Hebb's rule is successfully realized through optical experiments. The dual-light stimulation is associated with the change of the resin's fluorescence spectrum, thereby realizing the automatic generation of weights, and being able to complete the typical artificial intelligence task-pattern recognition.

[0061] Step S30, collecting a target image, inputting the target image into the input layer of the target network for preprocessing to obtain a preprocessed image, inputting the preprocessed image into the optical weight self-learning layer of the target network for inference to obtain an inference result, inputting the inference result into the output layer of the target network for classification, and outputting the classification result of the target image.

[0062] The step of inputting the target image into the input layer of the target network for preprocessing to obtain a preprocessed image specifically includes:

[0063] Inputting the target image into the input layer of the target network for denoising, filtering and normalization to obtain first intermediate data;

[0064] Compressing the first intermediate data and removing irrelevant information to reduce the number of weights to be learned by the optical weight self-learning layer, thereby obtaining second intermediate data;

[0065] The size of the second intermediate data is scaled according to the size of the optical weight self-learning layer to obtain third intermediate data, and initial feature extraction is performed on the third intermediate data to obtain the preprocessed image.

[0066] In this embodiment, after acquiring the target image, the target image is first input into the input layer of the target network for preliminary data preprocessing: image data is quickly preprocessed (such as noise reduction, filtering, normalization), irrelevant information is compressed, and the number of weights that need to be learned by the subsequent optical module is reduced. The image is scaled to a size suitable for the optical weight self-learning module, and the initial features are extracted at the same time.

[0067] Furthermore, the inputting the preprocessed image into the optical weight self-learning layer of the target network for reasoning to obtain the reasoning result specifically includes:

[0068] Inputting the preprocessed image into a weighted optical weight self-learning layer, super-projecting the preprocessed image through the rays emitted by the UV projection module, and obtaining a green fluorescence signal;

[0069] The fluorescent signal is converted into an electrical signal, and an inference result is obtained according to the electrical signal.

[0070] It is understandable that after the photonic neural network is trained, the optical module of the target network now carries the information weight, and the calculation can be completed by irradiating it with UV light. This process reflects the speed advantage of optical computing. When UV light irradiates the optical module with weight, the image information appears in the form of green fluorescence, and the green fluorescence signal is collected and converted into an electrical signal to be transmitted to the edge computing platform.

[0071] like Figure 4 As shown in the figure, when only 365nm light is irradiated on DCPI resin, red fluorescence is excited. When only 532nm light is irradiated on DCPI resin, green fluorescence is excited. When 365nm and 532nm light are irradiated on DCPI resin at the same time, the fluorescence spectrum of DCPI resin changes. At this time, 365nm light is used alone to irradiate the DCPI resin after the fluorescence spectrum change, and green fluorescence is excited.

[0072] Furthermore, the edge computing platform can perform further logical operations, such as classification and prediction, based on the characteristic data generated by the optical module to improve the reliability of the calculation. In addition to utilizing the speed advantage of optical computing, the processing combined with edge computing also combines the confidentiality of edge computing. For example, on the product line, the product photo is highly confidential. The characteristic information of the product photo is locked into the DCPI resin of the present invention, and then the DCPI resin with this characteristic is distributed to each production line. The edge computing platform then performs subsequent operations based on the characteristic data (excited fluorescence intensity) generated by the optical module (DCPI), and all calculation results are stored locally without uploading to the cloud; the combination of optical weight self-learning and edge computing real-time can greatly improve task processing efficiency and enhance data security through local processing and encryption technology. Whether in industry, medicine or other fields, this architecture can meet the needs of both efficiency and privacy protection.

[0073] It can be seen that the present invention constructs a photonic neural network based on DCPI resin, and the photonic neural network includes an input layer, an optical weight self-learning layer and an output layer; the photonic neural network is trained to obtain a target network; a target image is collected, and the target image is input into the input layer of the target network for preprocessing to obtain a preprocessed image, and the preprocessed image is input into the optical weight self-learning layer of the target network for reasoning to obtain a reasoning result, and the reasoning result is input into the output layer of the target network for classification, and the classification result of the target image is output. The photonic neural network based on DCPI resin in the present invention is an independent photonic neural network that does not rely on artificial neural networks. The generation of weights does not require a large amount of computing resources and energy consumption, and is an important step in human-like artificial intelligence.

[0074] Furthermore, if Figure 5 As shown, based on the above-mentioned image classification method based on DCPI resin, the present invention also provides an image classification method system based on DCPI resin, wherein the image classification method system based on DCPI resin includes:

[0075] A model building module 51 is used to build a photonic neural network based on DCPI resin, wherein the photonic neural network includes an input layer, an optical weight self-learning layer and an output layer;

[0076] A model training module 52 is used to train the photonic neural network to obtain a target network;

[0077] The optical computing module 53 is used to capture a target image, input the target image into the input layer of the target network for preprocessing to obtain a preprocessed image, input the preprocessed image into the optical weight self-learning layer of the target network for inference to obtain an inference result, input the inference result into the output layer of the target network for classification, and output the classification result of the target image.

[0078] Furthermore, if Figure 6 As shown, based on the above-mentioned DCPI resin-based image classification method and system, the present invention also provides a terminal accordingly, and the terminal includes a processor 10, a memory 20 and a display 30. Figure 6 Only some components of the terminal are shown, but it should be understood that it is not required to implement all of the components shown, and more or fewer components may be implemented instead.

[0079] In some embodiments, the memory 20 may be an internal storage unit of the terminal, such as a hard disk or memory of the terminal. In other embodiments, the memory 20 may also be an external storage device of the terminal, such as a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (SecureDigital, SD) card, a flash card (Flash Card), etc. equipped on the terminal. Further, the memory 20 may also include both an internal storage unit of the terminal and an external storage device. The memory 20 is used to store application software and various types of data installed in the terminal, such as the program code of the installation terminal, etc. The memory 20 may also be used to temporarily store data that has been output or is to be output. In one embodiment, a DCPI resin-based image classification method program 40 is stored on the memory 20, and the DCPI resin-based image classification method program 40 can be executed by the processor 10, thereby realizing the DCPI resin-based image classification method in the present application.

[0080] In some embodiments, the processor 10 may be a central processing unit (CPU), a microprocessor or other data processing chip, which is used to run the program code or process data stored in the memory 20, such as executing the DCPI resin-based image classification method.

[0081] In some embodiments, the display 30 may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, an OLED (Organic Light-Emitting Diode) touch device, etc. The display 30 is used to display information on the terminal and to display a visual user interface. The components of the terminal communicate with each other via a system bus.

[0082] In one embodiment, when the processor 10 executes the DCPI resin-based image classification method program 40 in the memory 20, the following steps are implemented:

[0083] Constructing a photonic neural network based on DCPI resin, wherein the photonic neural network includes an input layer, an optical weight self-learning layer and an output layer;

[0084] Training the photonic neural network to obtain a target network;

[0085] Capture a target image, input the target image into the input layer of the target network for preprocessing to obtain a preprocessed image, input the preprocessed image into the optical weight self-learning layer of the target network for inference to obtain an inference result, input the inference result into the output layer of the target network for classification, and output the classification result of the target image.

[0086] Wherein, the optical weight self-learning layer includes a UV projection module, a first microlens array module, a DCPI resin module, a second microlens array module and a laser irradiation module;

[0087] Wherein, the DCPI resin module is located between the first microlens array module and the second microlens array module.

[0088] Wherein, the DCPI resin module is obtained by imitating the biological learning process.

[0089] The step of training the photonic neural network to obtain a target network specifically includes:

[0090] Acquire a sample image, and preprocess the sample image to obtain a preprocessed sample image;

[0091] The preprocessed sample image is input into the UV projection module of the optical weight self-learning layer, and the preprocessed sample image is irradiated onto the DCPI resin module after passing through the first microlens array module by the first wavelength laser emitted by the UV projection module to obtain a first stimulation signal;

[0092] The second wavelength laser emitted by the laser irradiation module passes through the second microlens array module and then irradiates the DCPI resin module to obtain a second stimulation signal;

[0093] The first stimulation signal and the second stimulation signal are associated with each other through the DCPI resin module, and the weight of the information carried on the image is automatically generated, and the target network is obtained according to the weight.

[0094] The step of inputting the target image to the input layer of the target network for preprocessing to obtain a preprocessed image specifically includes:

[0095] Inputting the target image into the input layer of the target network for denoising, filtering and normalization to obtain first intermediate data;

[0096] Compressing the first intermediate data and removing irrelevant information to reduce the number of weights to be learned by the optical weight self-learning layer, thereby obtaining second intermediate data;

[0097] The size of the second intermediate data is scaled according to the size of the optical weight self-learning layer to obtain third intermediate data, and initial feature extraction is performed on the third intermediate data to obtain the preprocessed image.

[0098] The step of inputting the preprocessed image into the optical weight self-learning layer of the target network for inference to obtain an inference result specifically includes:

[0099] Inputting the preprocessed image into a weighted optical weight self-learning layer, super-projecting the preprocessed image through the rays emitted by the UV projection module, and obtaining a green fluorescence signal;

[0100] The fluorescent signal is converted into an electrical signal, and an inference result is obtained according to the electrical signal.

[0101] The first wavelength laser emitted by the UV projection module is a violet laser with a wavelength of 365 nm, and the second wavelength laser emitted by the laser irradiation module is a green laser with a wavelength of 532 nm.

[0102] The present invention also provides a computer-readable storage medium, wherein the computer-readable storage medium stores a DCPI resin-based image classification method program, and when the DCPI resin-based image classification method program is executed by a processor, the steps of the DCPI resin-based image classification method as described above are implemented.

[0103] In summary, the present invention proposes an image classification method, system and terminal based on DCPI resin, the method comprising: constructing a photon neural network based on DCPI resin, the photon neural network comprising an input layer, an optical weight self-learning layer and an output layer; training the photon neural network to obtain a target network; collecting a target image, inputting the target image into the input layer of the target network for preprocessing to obtain a preprocessed image, inputting the preprocessed image into the optical weight self-learning layer of the target network for reasoning to obtain a reasoning result, inputting the reasoning result into the output layer of the target network for classification, and outputting the classification result of the target image. The photon neural network based on DCPI resin in the present invention is an independent photon neural network that does not rely on artificial neural networks. The generation of weights does not require a large amount of computing resources and energy consumption, which is an important step in human-like artificial intelligence.

[0104] It should be noted that, in this article, the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or terminal including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or terminal. In the absence of further restrictions, an element defined by the sentence "comprises a ..." does not exclude the presence of other identical elements in the process, method, article or terminal including the element.

[0105] Of course, those skilled in the art can understand that all or part of the processes in the above-mentioned embodiments can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided by the present invention can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0106] It should be understood that the application of the present invention is not limited to the above examples. For ordinary technicians in this field, improvements or changes can be made based on the above description. All these improvements and changes should fall within the scope of protection of the claims attached to the present invention.

Claims

1. An image classification method based on DCPI resin, characterized in that, The image classification method based on DCPI resin comprises: Constructing a photonic neural network based on DCPI resin, wherein the photonic neural network includes an input layer, an optical weight self-learning layer and an output layer; Training the photonic neural network to obtain a target network; Capture a target image, input the target image into the input layer of the target network for preprocessing to obtain a preprocessed image, input the preprocessed image into the optical weight self-learning layer of the target network for inference to obtain an inference result, input the inference result into the output layer of the target network for classification, and output the classification result of the target image.

2. The image classification method based on DCPI resin according to claim 1, characterized in that, The optical weight self-learning layer includes a UV projection module, a first microlens array module, a DCPI resin module, a second microlens array module and a laser irradiation module; Wherein, the DCPI resin module is located between the first microlens array module and the second microlens array module.

3. The image classification method based on DCPI resin according to claim 2, characterized in that, The DCPI resin module is obtained by imitating a biological learning process.

4. The image classification method based on DCPI resin according to claim 3, characterized in that, The training of the photonic neural network to obtain a target network specifically includes: Acquire a sample image, and preprocess the sample image to obtain a preprocessed sample image; The preprocessed sample image is input into the UV projection module of the optical weight self-learning layer, and the preprocessed sample image is irradiated onto the DCPI resin module after passing through the first microlens array module by the first wavelength laser emitted by the UV projection module to obtain a first stimulation signal; The second wavelength laser emitted by the laser irradiation module passes through the second microlens array module and then irradiates the DCPI resin module to obtain a second stimulation signal; The first stimulation signal and the second stimulation signal are associated with each other through the DCPI resin module, and the weight of the information carried on the image is automatically generated, and the target network is obtained according to the weight.

5. The image classification method based on DCPI resin according to claim 1, characterized in that, The step of inputting the target image to the input layer of the target network for preprocessing to obtain a preprocessed image specifically includes: Inputting the target image into the input layer of the target network for denoising, filtering and normalization to obtain first intermediate data; Compressing the first intermediate data and removing irrelevant information to reduce the number of weights to be learned by the optical weight self-learning layer, thereby obtaining second intermediate data; The size of the second intermediate data is scaled according to the size of the optical weight self-learning layer to obtain third intermediate data, and initial feature extraction is performed on the third intermediate data to obtain the preprocessed image.

6. The image classification method based on DCPI resin according to claim 2, characterized in that, The step of inputting the preprocessed image into the optical weight self-learning layer of the target network for inference to obtain an inference result specifically includes: Inputting the preprocessed image into a weighted optical weight self-learning layer, super-projecting the preprocessed image through the rays emitted by the UV projection module, and obtaining a green fluorescence signal; The fluorescent signal is converted into an electrical signal, and an inference result is obtained according to the electrical signal.

7. The image classification method based on DCPI resin according to claim 2, characterized in that, The first wavelength laser emitted by the UV projection module is a violet laser with a wavelength of 365 nm, and the second wavelength laser emitted by the laser irradiation module is a green laser with a wavelength of 532 nm.

8. An image classification method system based on DCPI resin, characterized in that: The DCPI resin-based image classification method system comprises: A model building module, used to build a photonic neural network based on DCPI resin, wherein the photonic neural network includes an input layer, an optical weight self-learning layer and an output layer; A model training module, used to train the photonic neural network to obtain a target network; The optical computing module is used to capture a target image, input the target image into the input layer of the target network for preprocessing to obtain a preprocessed image, input the preprocessed image into the optical weight self-learning layer of the target network for inference to obtain an inference result, input the inference result into the output layer of the target network for classification, and output the classification result of the target image.

9. A terminal, characterized in that: The terminal includes: a memory, a processor, and a DCPI resin-based image classification method program stored in the memory and executable on the processor. When the DCPI resin-based image classification method program is executed by the processor, the steps of the DCPI resin-based image classification method as described in any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a DCPI resin-based image classification method program, and when the DCPI resin-based image classification method program is executed by a processor, the steps of the DCPI resin-based image classification method according to any one of claims 1 to 7 are implemented.

Images