Integrated visual intelligence platform based on local deployment of brain-like spiking neural network

Abstract

The application discloses a kind of local deployment integrated visual intelligence platform based on brain-like pulse neural network, belong to computer vision and artificial intelligence technical field, the platform includes user management module, project management module, image annotation module, model training module, model inference module and statistical representation module;Among them, user management module is used to realize the registration and login of user;Project management module is used to realize data loading, pre-processing and format conversion, and support model selection and design network structure;Statistical representation module is used to report generation and result visualization of image analysis result;Model training module and model inference module simultaneously support two kinds of technical routes of convolutional neural network and brain-like pulse neural network.The application can reduce the technical threshold of user on image analysis, through zero code development environment, so that non-professional person can easily carry out image analysis, without programming skill.

Description

Technical Field

[0001] This invention relates to the fields of computer vision and artificial intelligence, and in particular to an integrated visual intelligence platform based on a locally deployed neuromorphic spiking neural network. Background Technology

[0002] Neuromorphic computing is an emerging research direction in the field of artificial intelligence, aiming to simulate the working principle of the human brain to build more efficient and low-power intelligent computing systems. Spiking Neural Networks (SNNs), as one of the core technologies of neuromorphic computing, process information by simulating the spiking mechanism of neurons, exhibiting strong spatiotemporal information processing capabilities, high energy efficiency, and strong biological rationality. Compared with traditional artificial neural networks (ANNs), SNNs adopt an event-driven computing approach, firing spiking only when the input received by a neuron exceeds a threshold, thus demonstrating higher computational efficiency when processing sparse and asynchronous spatiotemporal data. In recent years, with the development of neuromorphic hardware, SNNs have shown enormous application potential in image recognition, object detection, signal processing, and other fields, particularly demonstrating significant advantages in edge computing and low-power scenarios.

[0003] Data annotation is a crucial part of image analysis, providing the learning materials needed for models. Early annotation relied primarily on manual work, which was inefficient and prone to errors. With advancements in image analysis technology, semi-automated annotation tools and intelligent annotation techniques have emerged, such as interactive segmentation tools, weakly supervised learning, and active learning, significantly improving annotation efficiency and accuracy. However, data annotation still faces challenges such as large datasets, inconsistent annotation quality, and high costs.

[0004] Model training and inference are the core components of image analysis, determining the final outcome of the task. Model training refers to training a deep learning model using labeled data, enabling it to recognize and segment objects in an image. Model inference refers to using the trained model to recognize and segment new images, obtaining the final analysis results. The development of model training and inference techniques has shifted from traditional methods to deep learning methods. Traditional methods primarily employ hand-designed feature extractors and classifiers, such as SIFT and SURF, but their performance is limited. The emergence of deep learning methods, such as Convolutional Neural Networks (CNNs), has achieved breakthroughs in image analysis, significantly improving the accuracy of recognition and segmentation. To further improve model performance, researchers have developed various model optimization methods, such as transfer learning, model compression, and model acceleration. However, traditional deep learning models typically require significant computational resources and energy, posing challenges for local deployment and application on edge devices.

[0005] In the current field of image analysis, there are various software platforms such as PaddleSeg, CVAT, Avizo, and MIPAR.

[0006] PaddleSeg is a deep learning-based image segmentation toolkit with powerful features, an active community, and scalability. It supports various image segmentation models and pre-trained models, allowing users to choose the appropriate model for training and inference based on their needs. PaddleSeg also boasts an active community where users can obtain help and support and contribute. Furthermore, PaddleSeg supports custom models and algorithms to meet specific user requirements, such as developing new loss functions and optimizing model structures. However, PaddleSeg is primarily geared towards algorithm developers and requires a certain level of programming knowledge, making it relatively difficult for non-computer science users to use. It also does not support neuromorphic spiking neural networks.

[0007] CVAT is a powerful web-based annotation tool that supports various annotation tasks, including image segmentation, object detection, and tracking. CVAT's web interface facilitates remote collaboration, improving annotation efficiency. Furthermore, CVAT supports multiple annotation formats, allowing users to easily export annotated data for model training. However, CVAT primarily focuses on data annotation and lacks complete support for model training and inference processes, as well as neuromorphic computing capabilities.

[0008] Avizo is a 3D and multidimensional data analysis platform with powerful data analysis tools suitable for complex research scenarios in fields such as materials science and geology. Avizo's 3D visualization capabilities help users intuitively analyze data and discover patterns and trends. Furthermore, Avizo supports various data processing and analysis tools, such as filtering, segmentation, and measurement, facilitating data analysis and modeling. However, Avizo is primarily based on traditional image processing methods, has limited support for deep learning and neuromorphic computing, and is expensive, resulting in high costs for local deployment.

[0009] MIPAR is an image analysis platform for materials science. It features innovative image analysis tools for studying the microscopic properties, microstructure, and performance relationships of materials. MIPAR's image segmentation and feature extraction tools help users quickly and accurately analyze material images and gain valuable insights. Furthermore, MIPAR supports various data visualization tools, facilitating the presentation and analysis of data results. However, MIPAR is primarily based on traditional algorithms, with insufficient support for deep learning and neuromorphic neural networks, and lacks comprehensive model training and customization capabilities.

[0010] In the field of neuromorphic computing, although some SNN frameworks and tools exist, such as SpykeTorch and BindsNET, these tools are mainly geared towards algorithm research and lack complete platform support for industry applications, making them difficult to directly apply to practical image analysis tasks in fields such as materials and medicine. Products that combine neuromorphic spiking neural networks with industry-specific visual analysis platforms are currently relatively rare.

[0011] With the growth of the core artificial intelligence industry, the demand for high-precision image analysis is increasing. In fields such as medicine, materials science, industrial manufacturing, agriculture, and security monitoring, high-precision image analysis has become a key factor in improving decision-making quality, optimizing operational processes, and enhancing product quality. For example, in the medical field, computer vision is becoming a crucial technology for improving diagnostic accuracy. It assists doctors in disease diagnosis by accurately analyzing medical images, such as identifying tumor boundaries and assessing lesions, and also shows great potential in surgical navigation and robot-assisted surgery. In pathology, it helps pathologists make more accurate cancer staging and treatment decisions by analyzing microscopic images. However, when faced with large-scale data analysis, especially high-resolution medical images, the accuracy and versatility of traditional image processing methods remain insufficient. In industrial manufacturing, computer vision is crucial for automated quality control and defect detection, improving production efficiency and product quality. In the automotive and electronics industries, it can detect welding, painting, and assembly defects, as well as problems on tiny circuit boards. However, these traditional methods are limited in high-speed production environments due to insufficient processing speed and accuracy.

[0012] Meanwhile, existing deep learning platforms are typically based on traditional artificial neural networks, which suffer from high computational demands, high energy consumption, and demanding hardware requirements, posing numerous challenges in local offline deployment scenarios. This is particularly true for users in fields such as materials science and medicine, who often need to process sensitive data locally and have high requirements for data security and privacy. However, local server computing resources are limited, and traditional deep learning models suffer from slow inference speeds and high resource consumption, impacting user experience.

[0013] Neuromorphic spiking neural networks, as a next-generation artificial intelligence technology, offer advantages such as low power consumption, high spatiotemporal resolution, and event-driven operation, making them highly suitable for local deployment and edge computing scenarios. However, there is currently a lack of complete platforms that combine neuromorphic technology with industry visual analysis needs, making it difficult for non-professional users to solve practical problems using neuromorphic technology. Summary of the Invention

[0014] This invention provides a locally deployed integrated visual intelligence platform based on a brain-like spiking neural network to solve the aforementioned technical problems existing in the prior art.

[0015] To solve the above-mentioned technical problems, the present invention provides the following technical solution: A locally deployed, integrated visual intelligence platform based on a neuromorphic spiking neural network includes: a user management module, a project management module, an image annotation module, a model training module, a model inference module, and a statistical representation module; wherein, The user management module is used to implement user registration and login; The project management module is used to load, preprocess, and convert the data related to the project to be processed, so as to ensure that the data can be adapted to subsequent model training and model inference; and supports the selection of model algorithms and the design of network structures, thereby providing support for the dynamic adjustment and deployment of models. The image annotation module is used for annotation operations on graphics; The model training module is used to train models to create new models and adjust old models; The model inference module is used to perform intelligent analysis of the target image using the selected model; The statistical characterization module is used for generating reports and visualizing image analysis results; The model training module and the model inference module support both convolutional neural networks and neuromorphic spiking neural networks, allowing users to flexibly choose different models according to their needs.

[0016] Furthermore, the platform also includes an automated convolutional neural network to spiking neural network conversion tool, which is used to perform model conversion while maintaining model accuracy, so that users can convert existing CNN models into SNN models as needed.

[0017] Furthermore, the platform supports multiple pulse coding methods to convert image data into pulse sequences, adapting to different SNN models and application scenarios.

[0018] Furthermore, the platform also includes a zero-code interactive intelligent annotation tool; The zero-code interactive intelligent annotation tool combines image processing technology and deep learning technology to automatically identify targets in images and generate annotations.

[0019] Furthermore, the platform supports local offline deployment, placing all data processing and storage on a local server.

[0020] Furthermore, the platform's data source combines public datasets with internally customized datasets to provide data support for basic training and testing; among them, public datasets provide a foundation for the general validation of the model, while internally customized datasets support performance optimization for specific application scenarios.

[0021] In another aspect, the present invention also provides an electronic device comprising a processor and a memory; wherein the memory stores at least one instruction, which is loaded and executed by the processor to implement the aforementioned platform.

[0022] In another aspect, the present invention also provides a computer-readable storage medium storing at least one instruction, which is loaded and executed by a processor to implement the above-described platform.

[0023] The technical solution provided by this invention integrates traditional deep learning and neuromorphic spiking neural network technologies, offering a highly accurate and automated image recognition and segmentation service platform. The platform lowers the technical barrier for image analysis for professionals in the fields of medicine, materials science, and industrial manufacturing. Its zero-code development environment allows non-professionals to easily perform image analysis without programming skills. The platform's data project management and model management functions, as well as its support for offline deployment, ensure the efficiency and security of data processing. In particular, the introduction of neuromorphic spiking neural networks provides users with more efficient and low-power inference options, better meeting the needs of local deployment scenarios. Its beneficial effects include at least the following: 1. Enhancing Industry Intelligence: Locally deployed visual intelligence computing platforms, by integrating advanced technologies such as deep learning, image segmentation, and object detection, significantly improve the data processing capabilities of various industries. The application of this technology not only promotes the popularization and application of artificial intelligence but also drives the overall improvement of societal intelligence, particularly in industries such as healthcare, security, and transportation, where enhanced intelligence is crucial for improving service quality and efficiency.

[0024] 2. Enhance market competitiveness: By utilizing intelligent visual computing platforms, enterprises can improve the quality of their products and services, thereby enhancing their market competitiveness. The application of this technology can help enterprises explore new markets, create more business opportunities, and thus increase revenue and market share.

[0025] 3. Facilitating Digital Transformation: The application of intelligent vision technology has helped traditional industries improve productivity and efficiency, reduce manual intervention, and drive the digital and automated transformation of these industries. This transformation not only enhances the competitiveness of enterprises but also provides new impetus for the upgrading of the entire industry.

[0026] 4. Promote Technological Innovation: The platform's application will drive the use of artificial intelligence (AI) technology across various fields, promoting its innovative development. This innovation can not only improve the efficiency and accuracy of existing technologies but also spawn new technologies and applications, thus propelling technological progress.

[0027] 5. Improved Local Deployment Performance and Energy Efficiency: The event-driven nature of neuromorphic spiking neural networks makes model inference more efficient, significantly improving inference speed and reducing computational resource consumption and energy consumption in local deployment scenarios. This is of great value to users who need to process large amounts of data locally, enabling them to achieve a better performance experience with limited hardware resources while reducing operating costs. Attached Figure Description

[0028] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0029] Figure 1 This is a structural block diagram of a locally deployed integrated visual intelligence platform based on a brain-like spiking neural network provided in an embodiment of the present invention. Figure 2 This is a system block diagram of the electronic device provided in the embodiments of the present invention. Detailed Implementation

[0030] To make the objectives, technical solutions, and advantages of the present invention clearer, the embodiments of the present invention will be described in further detail below with reference to the accompanying drawings.

[0031] First, it should be noted that in the embodiments of the present invention, the words "exemplarily," "for example," etc., are used to indicate that they are examples, illustrations, or descriptions. Any embodiment or design scheme described as "exemplary" in the present invention should not be construed as being more preferred or advantageous than other embodiments or design schemes. Specifically, the use of the term "exemplarily" is intended to present the concept in a specific manner. Furthermore, in the embodiments of the present invention, the meaning expressed by "and / or" can be both, or it can be either one or the other.

[0032] First Embodiment

[0033] To address the demand for detailed image analysis in fields such as materials science, biomedicine, and industrial manufacturing, this embodiment provides WisdomStore, a locally deployed, integrated visual intelligence platform based on neuromorphic spiking neural networks and traditional deep learning. This platform enables intelligent image analysis and supports local offline deployment and zero-code operation. Figure 1As shown, based on user needs analysis and software positioning, to ensure the final platform meets user requirements, including user management, data warehousing, image annotation, model inference, training, and statistical representation, the overall functional module structure of this embodiment is divided into six parts: user management, project management, image annotation, model training, model inference, and statistical representation. The model training and inference modules simultaneously support both traditional convolutional neural networks and neuromorphic spiking neural networks, allowing users to flexibly choose according to their needs.

[0034] Furthermore, the intelligent platform in this embodiment covers the complete technology stack from the underlying hardware to the user interface. It is divided into four main layers: 1) User Interface Layer: This layer is the top layer of the system and primarily handles user interaction. It comprises four main modules: data annotation interface, model training interface, model inference interface, and results display interface. Its main function is to provide an easy-to-use interface, enabling users to efficiently complete tasks such as data uploading, feature extraction, report generation, and results visualization. By designing an intuitive and user-friendly interface, this layer significantly lowers the technical barrier to entry, allowing even non-professional users to easily operate the system. This user-oriented design approach aligns with the trend in modern information system development, which seeks a balance between functionality and ease of use.

[0035] 2) Basic Service Layer: As the core computing component of the system, the basic service layer comprises two main modules: data processing and model management. The data processing module's primary tasks are data loading, preprocessing, and format conversion, ensuring the raw data is suitable for subsequent model training. The model management module supports model training, algorithm selection, and network structure design, covering key tasks such as model loading, configuration, segmentation, and result saving, providing support for dynamic model adjustment and deployment. Furthermore, the core algorithm section integrates various cutting-edge algorithms, such as Grabcut, Graphcut, and RITM, providing computational support for intelligent image annotation and model training inference.

[0036] 3) Language Framework Layer: This layer is developed based on the Python language and its ecosystem, supplemented by the PyQt5 framework. Python, with its extensive library support and flexibility, has become the preferred programming language for data processing and algorithm development, while PyQt5 simplifies graphical interface development, making user interface construction more efficient. Furthermore, the server-side is also developed based on Python. This unified technology stack design not only simplifies system maintenance but also ensures efficient collaboration between different modules.

[0037] 4) Underlying Service Layer: The data source combines public datasets (such as Berkeley, Grabcut, LVIS, COCO, etc.) with internally customized datasets, providing rich data support for basic training and testing. Public datasets provide the foundation for general model validation, while internal datasets support performance optimization for specific application scenarios. The system hardware configuration can meet the high computational demands of large-scale data processing and complex model training, ensuring the reliability and efficiency of system performance.

[0038] This platform features a clearly layered and modular architecture. Through hierarchical design, it constructs a highly integrated architecture, from underlying hardware support to high-level user functions, embodying a design philosophy of modularity, flexibility, and performance optimization. Each layer has clearly defined responsibilities, operating independently yet closely collaboratively, achieving full-process coverage from data annotation to result generation. This architectural design not only adapts to modern information processing needs but also provides ample flexibility for future functional expansion and system optimization. Specific implementations cover graphical interface design, data annotation, model training, and inference. By integrating multiple core algorithms and providing rich functionality, it efficiently supports the entire process from data acquisition to model application. It offers a convenient, efficient, and accurate interactive experience for non-computer professionals in industries such as materials science and medicine who require data annotation, helping users quickly complete various image processing tasks and enabling them to benefit from the rapid development of artificial intelligence.

[0039] In summary, this embodiment combines traditional deep learning with neuromorphic spiking neural networks (SNNs) to construct a unified visual intelligence analysis platform that supports model training, inference, and application using both CNN and SNN technologies. It achieves an organic integration of neuromorphic technology and industry-specific visual analysis needs. The platform allows users to train and infer models according to specific tasks and requirements, providing various basic models and customization options. Users can select appropriate model architectures, adjust parameters, and configure thresholds to optimize model performance and adaptability. Furthermore, the platform provides automated tools for converting convolutional neural networks (CNNs) to spiking neural networks, including steps such as weight normalization, threshold adjustment, and bias processing, achieving efficient model conversion while maintaining model accuracy. Users can easily convert existing CNN models to SNN models, enjoying the efficiency improvements brought by neuromorphic computing. In addition, the platform supports various pulse coding methods (rate coding, time coding, phase coding, etc.) to convert traditional image data into pulse sequences, adapting to different SNN models and application scenarios. The coding module is configurable and scalable, providing a flexible solution for input processing of neuromorphic models. Furthermore, the event-driven spiking neural network inference engine supports efficient spike propagation computation, as well as CPU / GPU acceleration and neuromorphic hardware integration, ensuring efficient inference in local deployment scenarios. Further, this platform offers a zero-code interactive intelligent annotation method: this method combines image processing and deep learning technologies to automatically identify targets in images and generate annotations, significantly improving annotation efficiency and accuracy. For example, click-based annotation, floating annotation, and frame-based annotation, as well as a multi-object detection and segmentation framework based on the SAM model, effectively address the shortcomings of existing annotation tools. Moreover, this platform supports local offline deployment, placing all data processing and storage on a local server, effectively ensuring data security and privacy. This is particularly suitable for fields involving sensitive data, such as healthcare and finance.

[0040] Second Embodiment

[0041] This embodiment provides an electronic device, such as... Figure 2 As shown, the electronic device includes a processor and a memory; wherein the processor and the memory can be connected via a communication bus; the memory stores at least one instruction, which is loaded and executed by the processor to implement the solution of the first embodiment described above. Furthermore, the electronic device may also include a transceiver, the processor and the transceiver can be connected via a communication bus, and the transceiver is used to communicate with other devices.

[0042] Below, in conjunction with Figure 2 A detailed introduction to each component of this electronic device is provided below: The processor is the control center of the electronic device. The electronic device may include multiple processors, each of which can be a single-core processor (single-CPU) or a multi-core processor (multi-CPU). The term "processor" can refer to a single processor or a collective term for multiple processing elements. For example, a processor can be one or more central processing units (CPUs), other general-purpose processors, application-specific integrated circuits (ASICs), or one or more integrated circuits configured to implement embodiments of the present invention, such as one or more digital signal processors (DSPs), one or more field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor can be a microprocessor or any conventional processor. The processor can perform various functions of the electronic device by running or executing software programs stored in memory and by calling data stored in memory.

[0043] In a specific implementation, as one example, the processor may include one or more CPUs, for example... Figure 2 CPU0 and CPU1 shown are, of course, merely illustrative examples.

[0044] The memory is used to store the software program that executes the solution of the present invention, and the processor controls its execution. For specific implementation methods, please refer to the above method embodiments, which will not be repeated here.

[0045] Optionally, the memory may be a read-only memory (ROM) or other type of static storage device capable of storing static information and instructions, random access memory (RAM) or other type of dynamic storage device capable of storing information and instructions, or electrically erasable programmable read-only memory (EEPROM), compact disc read-only memory (CD-ROM) or other optical disc storage, optical disc storage (including compressed optical discs, laser discs, optical discs, digital universal optical discs, Blu-ray discs, etc.), magnetic disk storage media or other magnetic storage devices, or any other medium capable of carrying or storing desired program code in the form of instructions or data structures and accessible by a computer, but not limited thereto. The memory may be integrated with the processor or exist independently, and may be accessed through the interface circuit of the electronic device ( Figure 2 (Not shown in the image) is coupled to the processor; however, this embodiment of the invention does not impose specific limitations on this.

[0046] The transceiver may include a receiver and a transmitter. Figure 2 (Not shown separately). The receiver is used to implement the receiving function, and the transmitter is used to implement the transmitting function. The transceiver can be integrated with the processor or exist independently, and can be connected through the interface circuit of the electronic device (…). Figure 2 (Not shown in the image) is coupled to the processor, and this embodiment of the invention does not specifically limit this.

[0047] In addition, it should be noted that, Figure 2 The structure of the electronic device shown is not intended to limit the device. Actual devices may include more or fewer components than shown, or combine certain components, or have different component arrangements. Furthermore, the technical effects achieved by this electronic device when performing the method of the first embodiment described above can be referenced to the technical effects described in the first embodiment; therefore, they will not be repeated here.

[0048] Third Embodiment

[0049] This embodiment provides a computer-readable storage medium storing at least one instruction, which is loaded and executed by a processor to implement the solution of the first embodiment described above. The computer-readable storage medium may be a ROM, random access memory, CD-ROM, magnetic tape, floppy disk, or optical data storage device, etc. The instruction stored therein can be loaded and executed by a processor in a terminal.

[0050] Furthermore, it should be noted that the present invention can be provided as a method, apparatus, or computer program product. Therefore, embodiments of the present invention can take the form of a completely or partially hardware embodiment, a completely or partially software embodiment, or an embodiment combining software and hardware aspects. Moreover, when implemented in software, embodiments of the present invention can take the form of a computer program product implemented on one or more computer-usable storage media containing computer-usable program code. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer program are loaded or executed on a computer, all or part of the processes or functions described in the embodiments of the present invention are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any usable medium accessible to a computer or a data storage device such as a server or data center containing one or more sets of usable media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium. A semiconductor medium can be a solid-state drive (SSD).

[0051] Embodiments of the present invention are described with reference to flowchart illustrations and / or block diagrams of methods, terminal devices (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, embedded processor, or other programmable data processing terminal device to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing terminal device, generate instructions for implementing the flowchart illustrations. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0052] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing terminal device to operate in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1The functions specified in one or more boxes. These computer program instructions may also be loaded onto a computer or other programmable data processing terminal equipment to cause a series of operational steps to be performed on the computer or other programmable terminal equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable terminal equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0053] It should also be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. The terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or terminal device that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or terminal device. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or terminal device that includes said element. Furthermore, the term "and / or" is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A alone, A and B simultaneously, and B alone, where A and B can be singular or plural. Additionally, the character " / " in this text generally indicates an "or" relationship between the preceding and following objects, but it can also indicate an "AND / OR" relationship. Please refer to the context for specific interpretations. "At least one" refers to one or more items, while "more than" refers to two or more items. "At least one of the following" or similar expressions refer to any combination of these items, including any combination of single or multiple items. For example, at least one of a, b, or c can be represented as: a, b, c, ab, ac, bc, or abc, where a, b, and c can be single or multiple.

[0054] Furthermore, it is understood that in various embodiments of the present invention, the order of the above-mentioned process numbers does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0055] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0056] In the several embodiments provided by this invention, it should be understood that the disclosed devices, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative. For instance, the division of functional modules / units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another device, or some features may be ignored or not executed. Furthermore, the shown or discussed mutual couplings or direct couplings or communication connections may be through some interfaces; indirect couplings or communication connections between devices or units may be electrical, mechanical, or other forms. Units described as separate components may or may not be physically separate, and components shown as units may or may not be physical units, i.e., they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs. Additionally, the functional units in the various embodiments of this invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0057] If the method is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0058] Finally, it should be noted that the above description is merely a preferred embodiment of the present invention. It should be pointed out that although preferred embodiments of the present invention have been described, those skilled in the art, once they understand the basic inventive concept of the present invention, can make several improvements and modifications without departing from the principles described herein. These improvements and modifications should also be considered within the scope of protection of the present invention. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of the embodiments of the present invention.

Claims

1. A locally deployed integrated visual intelligence platform based on a brain-like spiking neural network, characterized in that, The integrated visual intelligence platform includes: a user management module, a project management module, an image annotation module, a model training module, a model inference module, and a statistical representation module; among which, The user management module is used to implement user registration and login; The project management module is used to load, preprocess, and convert the data related to the project to be processed, so as to ensure that the data can be adapted to subsequent model training and model inference; and supports the selection of model algorithms and the design of network structures, thereby providing support for the dynamic adjustment and deployment of models. The image annotation module is used for annotation operations on graphics; The model training module is used to train models to create new models and adjust old models; The model inference module is used to perform intelligent analysis of the target image using the selected model; The statistical characterization module is used for generating reports and visualizing image analysis results; The model training module and the model inference module support both convolutional neural networks and neuromorphic spiking neural networks, allowing users to flexibly choose different models according to their needs.

2. The locally deployed integrated visual intelligence platform based on a neuromorphic spiking neural network as described in claim 1, characterized in that, The platform also includes an automated tool for converting convolutional neural networks to spiking neural networks. This tool is used to convert models while maintaining their accuracy, allowing users to convert existing CNN models into SNN models as needed.

3. The locally deployed integrated visual intelligence platform based on a neuromorphic spiking neural network as described in claim 1, characterized in that, The platform supports multiple pulse coding methods to convert image data into pulse sequences, adapting to different SNN models and application scenarios.

4. The locally deployed integrated visual intelligence platform based on a neuromorphic spiking neural network as described in claim 1, characterized in that, The platform also includes a zero-code interactive intelligent annotation tool; The zero-code interactive intelligent annotation tool combines image processing technology and deep learning technology to automatically identify targets in images and generate annotations.

5. The locally deployed integrated visual intelligence platform based on a neuromorphic spiking neural network as described in claim 1, characterized in that, The platform supports local offline deployment, placing all data processing and storage on a local server.

6. The locally deployed integrated visual intelligence platform based on a neuromorphic spiking neural network as described in claim 1, characterized in that, The platform's data source combines public datasets and internally customized datasets to provide data support for basic training and testing. Public datasets provide a foundation for validating the generality of the model, while internally customized datasets support performance optimization for specific application scenarios.

Images