Cloud-based image data management system

Abstract

This invention relates to a cloud-based image data management system, comprising a cloud service node, a content acquisition device, a model conversion device, an encoding prediction mechanism, and a parameter display mechanism. The cloud-based image data management system of this invention has wide applications and intelligent operation.

Description

Technical Field

[0001] This invention relates to the field of cloud computing, and more particularly to a cloud-based image data management system. Background Technology

[0002] There are several types of cloud computing, mainly including public cloud, private cloud, and hybrid cloud.

[0003] Public cloud: Owned and operated by a third-party cloud service provider, it provides computing resources via the internet, such as Microsoft Azure. The hardware, software, and other supporting infrastructure in a public cloud are managed and maintained by the cloud provider.

[0004] Private cloud: Dedicated to the use of a single enterprise or organization, it can reside above the company's on-site data center or be hosted by a third-party service provider. Private clouds maintain services and infrastructure on a dedicated network.

[0005] Hybrid cloud: combines public and private clouds, allowing data and applications to be shared between the two, providing more flexible business processing and optimizing existing infrastructure, security, and compliance.

[0006] CN118470340A relates to the field of image processing technology, specifically a cloud-based image processing method, apparatus, storage medium, and electronic device. The method includes: outputting a corresponding bitemporal difference fusion image; extracting features from the bitemporal difference fusion image to construct a difference image; obtaining corresponding univariate and binary potential functions using clustering algorithms and Gaussian functions; and combining the univariate and binary potential functions to calculate the posterior probability of pixel changes. It generates a corresponding bitemporal difference fusion image from multiple bitemporal image data, accelerating parallel data reading. By using clustering algorithms and Gaussian functions to obtain corresponding univariate and binary potential functions, it optimizes the computation process in a distributed environment, reducing redundant calculations. By combining the univariate and binary potential functions and using a mean-field approximation algorithm to calculate the posterior probability of pixel changes, it achieves efficient pixel change detection while ensuring image processing accuracy.

[0007] CN111724451B relates to a cloud-based method, system, terminal, and storage medium for accelerating tomographic image reconstruction. It includes: acquiring raw scan data of an object and uploading the raw scan data to a cloud computing system; wherein the cloud computing system includes a Master and a set number of Workers; configuring a parallel computing environment on the cloud computing system using a parallel computing framework through the Master, dividing the tomographic image reconstruction task of the raw scan data into a set number of subtasks, assigning the subtasks to the set number of Workers for reconstruction calculations, and integrating the reconstruction calculation results of all Workers to generate a tomographic reconstructed image of the raw scan data. Embodiments of this application are compatible with multiple computing frameworks such as MapReduce and MPI, achieving cloud-based parallel computing, and are convenient to use, flexible in computation, low in cost, and not limited by geographical location.

[0008] CN118135376A discloses a cloud computing-based image data processing method and system, belonging to the field of image processing technology. The method involves pre-clustering initialization of the image to be processed, determining the starting position and initial distribution region based on the pre-clustering initialization results; deploying simulated particles of the ant colony algorithm in the initial distribution region and performing pheromone initialization to obtain an initial pheromone matrix; using the simulated particles of the ant colony algorithm to perform iterative optimization processing on the image to be processed, dividing the simulated particles into elite particles and ordinary particles, and updating the initial pheromone matrix to obtain an updated pheromone matrix; stopping the iteration when a preset improvement magnitude threshold is reached, and outputting the segmentation result of the image to be processed. This method combines heuristic initialization and an elite ant strategy to process the image, reducing blindness in the exploration process and reducing unnecessary exploration, thereby reducing the overall number of iterations and improving computational efficiency. Summary of the Invention

[0009] To overcome the technical problems in the prior art, this invention proposes a cloud-based image data management system. Based on cloud computing, for the same intra-frame coded block, a customized artificial intelligence model is used to intelligently predict the coding performance of other colors that have not yet undergone intra-frame coding based on the coding data of each color component that has been intra-coded. Thus, the coding performance of other colors can be determined before intra-frame coding is performed, creating space for adjusting coding parameters and improving the speed and performance of image data processing.

[0010] According to the present invention, a cloud computing-based image data management system is provided, the system comprising:

[0011] The cloud computing service node is used to obtain the red-green component image block, black-and-white component image block and yellow-blue component image block corresponding to the current image block. The red-green component image block and black-and-white component image block corresponding to the current image block are both intra-frame encoded image blocks, and the yellow-blue component image block corresponding to the current image block is an un-intra-frame encoded image block.

[0012] The content acquisition device is connected to the cloud computing service node and is used to acquire the red-green component image block and black-and-white component image block corresponding to the current image block, respectively, and to acquire the encoding speed, encoding computation amount and encoding compression ratio corresponding to the red-green component image block, and the encoding speed, encoding computation amount and encoding compression ratio corresponding to the black-and-white component image block.

[0013] A model conversion device is used to perform multi-level conversion processing on a deep neural network to obtain an intelligent prediction model. The multi-level conversion processing on the deep neural network to obtain an intelligent prediction model includes: each level of conversion processing performed on the deep neural network is a training action performed on the deep neural network, and the number of levels of conversion processing performed on the deep neural network is positively correlated with the number of pixels occupied by the image block.

[0014] The encoding prediction mechanism, connected to the content acquisition device, the model conversion device, and the cloud computing service node, is used to intelligently predict the encoding speed, encoding computation, and encoding compression ratio of the yellow-blue component image block based on the reference encoding block corresponding to the current image block, the yellow-blue component image block, the red-green component image block, the black-and-white component image block, the red-green component reconstruction block, the black-and-white component reconstruction block, the encoding speed, encoding computation, and encoding compression ratio of the red-green component image block, and the encoding speed, encoding computation, and encoding compression ratio of the black-and-white component image block.

[0015] A parameter display mechanism, connected to the encoding prediction mechanism, is used to receive and display the encoding speed, encoding computation amount, and encoding compression ratio corresponding to the yellow-blue component image blocks;

[0016] Specifically, the intelligent prediction model intelligently predicts the coding speed, coding computation, and coding compression ratio of the yellow-blue component image block, red-green component image block, black-and-white component image block, red-green component reconstruction block, black-and-white component reconstruction block, red-green component image block, and red-green component image block, as well as the coding speed, coding computation, and coding compression ratio of the black-and-white component image block, based on the corresponding yellow-blue component image block, red-green component image block, and black-and-white component image block of the current image block. This includes selecting the yellow-blue component image block, red-green component image block, and black-and-white component image block of the same reference coding block as their respective intra-frame coding reference frames.

[0017] Therefore, it can be seen that the present invention has at least the following three important inventive points:

[0018] Key Invention Point A: The invention employs a cloud computing service node to obtain the red-green component image block, black-and-white component image block, and yellow-and-blue component image block corresponding to the current image block. The red-green component image block and the black-and-white component image block corresponding to the current image block are both intra-coded image blocks, while the yellow-and-blue component image block corresponding to the current image block is an un-intra-coded image block. Furthermore, the invention obtains the red-green component reconstruction block and the black-and-white component reconstruction block corresponding to the red-green component image block and the black-and-white component image block, respectively. Simultaneously, it obtains the encoding speed, encoding computation, and encoding compression ratio corresponding to the red-green component image block, and the encoding speed, encoding computation, and encoding compression ratio corresponding to the black-and-white component image block. This provides foundational data for predicting the encoding performance of other colors that have not undergone intra-coded encoding.

[0019] Key Invention Point B: Performing multi-level transformation processing on a deep neural network to obtain an intelligent prediction model, wherein performing multi-level transformation processing on a deep neural network to obtain an intelligent prediction model includes: each level of transformation processing performed on the deep neural network is a training action performed on the deep neural network, and the number of levels of transformation processing performed on the deep neural network is positively correlated with the number of pixels occupied by the image block, thereby providing an artificial intelligence model for predicting the coding performance of other colors that have not undergone intra-frame coding.

[0020] Key Invention Point C: An intelligent prediction model is employed to intelligently predict the coding speed, coding computation, and coding compression ratio of the yellow-blue component image block, red-green component image block, black-and-white component image block, red-green component reconstruction block, black-and-white component reconstruction block, red-green component image block, and the coding speed, coding computation, and coding compression ratio of the black-and-white component image block, based on the reference coding block corresponding to the current image block, the yellow-blue component image block corresponding to the current image block, the yellow-blue component image block corresponding to the current image block, and the black-and-white component image block corresponding to the black-and-white component image block. This enables intelligent prediction of the coding performance of other colors that have not undergone intra-frame coding based on the coding data of each color component that has already been intra-coded.

[0021] The cloud-based image data management system of this invention has wide applications and intelligent operation. Because it can, based on cloud computing, intelligently predict the encoding performance of other colors that have not yet undergone intra-frame encoding by using a customized artificial intelligence model for the same intra-frame encoded block, based on the encoded data of each color component that has already been intra-coded, it improves the speed and performance of image data processing. Attached Figure Description

[0022] The embodiments of the present invention will now be described with reference to the accompanying drawings, wherein:

[0023] Figure 1 This is a block diagram illustrating the internal structure of a cloud-based image data management system according to embodiment 1 of the present invention. Detailed Implementation

[0024] Currently, cloud computing still has significant application potential. For example, in image data processing, for the same intra-coded block, it is desirable to intelligently predict the coding performance of other colors that have not yet undergone intra-coded based on the coding data of each color component that has already been intra-coded. This would allow the coding performance of other colors to be determined before intra-coded, creating space for adjusting coding parameters. However, existing technologies lack corresponding solutions that utilize cloud computing.

[0025] The implementation scheme of the cloud-based image data management system of the present invention will now be described in detail with reference to the accompanying drawings.

[0026] Figure 1 The diagram above shows the internal structure of a cloud-based image data management system according to embodiment 1 of the present invention. The system includes:

[0027] The cloud computing service node is used to obtain the red-green component image block, black-and-white component image block and yellow-blue component image block corresponding to the current image block. The red-green component image block and black-and-white component image block corresponding to the current image block are both intra-frame encoded image blocks, and the yellow-blue component image block corresponding to the current image block is an un-intra-frame encoded image block.

[0028] For example, a cloud computing service node is used to obtain the red-green component image block, black-and-white component image block, and yellow-blue component image block corresponding to the current image block. The red-green component image block and black-and-white component image block corresponding to the current image block are both intra-frame encoded image blocks. The yellow-blue component image block corresponding to the current image block is an un-intra-frame encoded image block, including: each image block has a square shape.

[0029] The content acquisition device is connected to the cloud computing service node and is used to acquire the red-green component image block and black-and-white component image block corresponding to the current image block, respectively, and to acquire the encoding speed, encoding computation amount and encoding compression ratio corresponding to the red-green component image block, and the encoding speed, encoding computation amount and encoding compression ratio corresponding to the black-and-white component image block.

[0030] A model conversion device is used to perform multi-level conversion processing on a deep neural network to obtain an intelligent prediction model. The multi-level conversion processing on the deep neural network to obtain an intelligent prediction model includes: each level of conversion processing performed on the deep neural network is a training action performed on the deep neural network, and the number of levels of conversion processing performed on the deep neural network is positively correlated with the number of pixels occupied by the image block.

[0031] The encoding prediction mechanism, connected to the content acquisition device, the model conversion device, and the cloud computing service node, is used to intelligently predict the encoding speed, encoding computation, and encoding compression ratio of the yellow-blue component image block based on the reference encoding block corresponding to the current image block, the yellow-blue component image block, the red-green component image block, the black-and-white component image block, the red-green component reconstruction block, the black-and-white component reconstruction block, the encoding speed, encoding computation, and encoding compression ratio of the red-green component image block, and the encoding speed, encoding computation, and encoding compression ratio of the black-and-white component image block.

[0032] A parameter display mechanism, connected to the encoding prediction mechanism, is used to receive and display the encoding speed, encoding computation amount, and encoding compression ratio corresponding to the yellow-blue component image blocks;

[0033] The intelligent prediction model intelligently predicts the coding speed, coding computation, and coding compression ratio of the yellow-blue component image block, red-green component image block, black-and-white component image block, red-green component reconstruction block, black-and-white component reconstruction block, red-green component image block, and black-and-white component image block corresponding to the current image block, as well as the coding speed, coding computation, and coding compression ratio corresponding to the black-and-white component image block. This includes: the yellow-blue component image block, red-green component image block, and black-and-white component image block corresponding to the current image block select the yellow-blue component image block, red-green component image block, and black-and-white component image block of the same reference coding block as their respective intra-frame coding reference frames.

[0034] The method of using an intelligent prediction model to intelligently predict the encoding speed, encoding computation, and encoding compression ratio of the yellow-blue component image block, red-green component image block, black-and-white component image block, red-green component reconstruction block, black-and-white component reconstruction block, red-green component image block, and the encoding speed, encoding computation, and encoding compression ratio of the black-and-white component image block also includes: using the pixel values ​​corresponding to each pixel point of each image block as input data to the intelligent prediction model;

[0035] In addition, the input data of the step of using an intelligent prediction model to intelligently predict the coding speed, coding computation, and coding compression ratio of the yellow-blue component image block, the red-green component image block, the black-and-white component image block, the red-green component reconstruction block, the black-and-white component reconstruction block, the coding speed, coding computation, and coding compression ratio of the red-green component image block, and the coding speed, coding computation, and coding compression ratio of the black-and-white component image block are subjected to numerical normalization processing to obtain the input data after numerical normalization processing.

[0036] The step of using an intelligent prediction model to intelligently predict the coding speed, coding computation, and coding compression ratio of the yellow-blue component image block, red-green component image block, black-and-white component image block, red-green component reconstruction block, black-and-white component reconstruction block, and the coding speed, coding computation, and coding compression ratio of the red-green component image block, and the coding speed, coding computation, and coding compression ratio of the black-and-white component image block, involves the input data undergoing numerical normalization to obtain the input data after numerical normalization. This includes: using... The intelligent prediction model intelligently predicts the coding speed, coding computation, and coding compression ratio of the yellow-blue component image block, the red-green component image block, the black-and-white component image block, the red-green component reconstruction block, the black-and-white component reconstruction block, the coding speed, coding computation, and coding compression ratio of the red-green component image block, and the coding speed, coding computation, and coding compression ratio of the black-and-white component image block. The input data for the step of intelligent prediction of the yellow-blue component image block is subjected to numerical normalization processing based on hexadecimal numerical conversion to obtain the input data after numerical normalization processing.

[0037] The step of using an intelligent prediction model to intelligently predict the coding speed, coding computation, and coding compression ratio of the yellow-blue component image block, red-green component image block, black-and-white component image block, red-green component reconstruction block, black-and-white component reconstruction block, and the coding speed, coding computation, and coding compression ratio of the red-green component image block, and the coding speed, coding computation, and coding compression ratio of the black-and-white component image block, involves the input data undergoing numerical normalization to obtain the input data after numerical normalization. This includes: [The text abruptly ends here, likely due to an incomplete sentence or a formatting error.] The input data for the step of intelligently predicting the coding speed, coding computation, and coding compression ratio of the yellow-blue component image block, red-green component image block, black-and-white component image block, red-green component reconstruction block, black-and-white component reconstruction block, coding speed, coding computation, and coding compression ratio of the red-green component image block, and the coding speed, coding computation, and coding compression ratio of the black-and-white component image block, based on the reference coding block corresponding to the current image block, is subjected to numerical normalization processing based on octal numerical conversion to obtain the input data after numerical normalization processing.

[0038] The step of using an intelligent prediction model to intelligently predict the coding speed, coding computation, and coding compression ratio of the yellow-blue component image block, red-green component image block, black-and-white component image block, red-green component reconstruction block, black-and-white component reconstruction block, and the coding speed, coding computation, and coding compression ratio of the red-green component image block, and the coding speed, coding computation, and coding compression ratio of the black-and-white component image block, involves the input data undergoing numerical normalization to obtain the input data after numerical normalization. This includes: [The text abruptly ends here, likely due to an incomplete sentence or a formatting error.] The input data for the step of intelligently predicting the coding speed, coding computation, and coding compression ratio of the yellow-blue component image block corresponding to the current image block, the yellow-blue component image block, the red-green component image block, the black-and-white component image block, the red-green component reconstruction block, the black-and-white component reconstruction block, the coding speed, coding computation, and coding compression ratio of the red-green component image block, and the coding speed, coding computation, and coding compression ratio of the black-and-white component image block are subjected to numerical normalization processing based on binary numerical conversion to obtain the input data after numerical normalization processing.

[0039] The step of using an intelligent prediction model to intelligently predict the coding speed, coding computation, and coding compression ratio of the yellow-blue component image block, red-green component image block, black-and-white component image block, red-green component reconstruction block, black-and-white component reconstruction block, and the coding speed, coding computation, and coding compression ratio of the red-green component image block, and the coding speed, coding computation, and coding compression ratio of the black-and-white component image block, involves the input data undergoing numerical normalization to obtain the input data after numerical normalization. This includes: [The text abruptly ends here, likely due to an incomplete sentence or a formatting error.] The input data for the step of using an intelligent prediction model to intelligently predict the coding speed, coding computation, and coding compression ratio of the yellow-blue component image block, red-green component image block, black-and-white component image block, red-green component reconstruction block, black-and-white component reconstruction block, coding speed, coding computation, and coding compression ratio of the red-green component image block, and the coding speed, coding computation, and coding compression ratio of the black-and-white component image block are subjected to numerical normalization processing based on decimal numerical conversion to obtain the input data after numerical normalization processing.

[0040] The output data of the step of using an intelligent prediction model to intelligently predict the coding speed, coding computation, and coding compression ratio of the yellow-blue component image block, red-green component image block, black-and-white component image block, red-green component reconstruction block, black-and-white component reconstruction block, coding speed, coding computation, and coding compression ratio of the red-green component image block, and coding speed, coding computation, and coding compression ratio of the black-and-white component image block are in a numerical representation after numerical normalization.

[0041] The output data of the step of intelligently predicting the coding speed, coding computation, and coding compression ratio of the yellow-blue component image block, red-green component image block, black-and-white component image block, red-green component reconstruction block, black-and-white component reconstruction block, and red-green component image block based on the reference coding block, yellow-blue component image block, red-green ... The output data for the steps of intelligent prediction of the encoding speed, encoding computation, and encoding compression ratio of the red-green component image blocks, black-and-white component image blocks, red-green component reconstruction blocks, black-and-white component reconstruction blocks, red-green component image blocks, and black-and-white component image blocks are one of the following: encoding speed, encoding computation, and encoding compression ratio.

[0042] Furthermore, in the cloud-based image data management system, a model conversion device is used to perform multi-level conversion processing on a deep neural network to obtain an intelligent prediction model. This multi-level conversion processing includes: each level of conversion processing performed on the deep neural network is a training action performed on the deep neural network; and the positive correlation between the number of levels of conversion processing performed on the deep neural network and the number of pixels occupied by an image block includes: using a content conversion function to represent the positive correlation between the number of levels of conversion processing performed on the deep neural network and the number of pixels occupied by an image block.

[0043] While specific embodiments of the invention have been described in detail, the breadth and scope of the invention should not be limited to the exemplary embodiments described above, but should be defined solely by the following claims and their equivalents. All modifications and alterations falling within the spirit of this invention are intended to be protected.

Claims

1. A cloud-based image data management system, characterized in that, The system includes: The cloud computing service node is used to obtain the red-green component image block, black-and-white component image block and yellow-blue component image block corresponding to the current image block. The red-green component image block and black-and-white component image block corresponding to the current image block are both intra-frame encoded image blocks, and the yellow-blue component image block corresponding to the current image block is an un-intra-frame encoded image block. The content acquisition device is connected to the cloud computing service node and is used to acquire the red-green component image block and black-and-white component image block corresponding to the current image block, respectively, and to acquire the encoding speed, encoding computation amount and encoding compression ratio corresponding to the red-green component image block, and the encoding speed, encoding computation amount and encoding compression ratio corresponding to the black-and-white component image block. A model conversion device is used to perform multi-level conversion processing on a deep neural network to obtain an intelligent prediction model. The multi-level conversion processing on the deep neural network to obtain an intelligent prediction model includes: each level of conversion processing performed on the deep neural network is a training action performed on the deep neural network, and the number of levels of conversion processing performed on the deep neural network is positively correlated with the number of pixels occupied by the image block. The encoding prediction mechanism, connected to the content acquisition device, the model conversion device, and the cloud computing service node, is used to intelligently predict the encoding speed, encoding computation, and encoding compression ratio of the yellow-blue component image block based on the reference encoding block corresponding to the current image block, the yellow-blue component image block, the red-green component image block, the black-and-white component image block, the red-green component reconstruction block, the black-and-white component reconstruction block, the encoding speed, encoding computation, and encoding compression ratio of the red-green component image block, and the encoding speed, encoding computation, and encoding compression ratio of the black-and-white component image block. This includes: the yellow-blue component image block, the red-green component image block, and the black-and-white component image block corresponding to the current image block each selecting the yellow-blue component image block, the red-green component image block, and the black-and-white component image block of the same reference encoding block as their respective intra-frame coding reference frames; The parameter display mechanism, connected to the encoding prediction mechanism, is used to receive and display the encoding speed, encoding computation amount, and encoding compression ratio corresponding to the yellow-blue component image blocks.

2. The cloud-based image data management system as described in claim 1, characterized in that: The intelligent prediction model intelligently predicts the encoding speed, encoding computation, and encoding compression ratio of the yellow-blue component image block, red-green component image block, black-and-white component image block, red-green component reconstruction block, black-and-white component reconstruction block, red-green component image block, and black-and-white component image block, based on the encoding speed, encoding computation, and encoding compression ratio of the yellow-blue component image block and the black-and-white component image block. It also includes using the pixel values ​​corresponding to each pixel point of each image block as input data to the intelligent prediction model.

3. The cloud-based image data management system as described in claim 2, characterized in that: The input data is subjected to numerical normalization to obtain the input data after numerical normalization.

4. The cloud-based image data management system as described in claim 3, characterized in that: The input data is subjected to numerical normalization processing based on hexadecimal, octal, binary, or decimal numerical conversion to obtain the input data after numerical normalization processing.

5. The cloud-based image data management system as described in claim 2, characterized in that: The intelligent prediction model intelligently predicts the coding speed, coding computation, and coding compression ratio of the yellow-blue component image block, red-green component image block, black-and-white component image block, red-green component reconstruction block, black-and-white component reconstruction block, coding speed, coding computation, and coding compression ratio of the red-green component image block, and the coding speed, coding computation, and coding compression ratio of the black-and-white component image block. The output data of this step is a numerically normalized representation.

6. The cloud-based image data management system as described in claim 5, characterized in that: The output data is a numerical representation after normalization processing based on hexadecimal, octal, binary, or decimal numerical conversion.

Images