Quantitative decomposition logical storage architecture for simplifying image recognition
By using alphabetical sorting and morphological quantization to decompose the logical storage architecture, the problems of hardware resource waste and privacy leakage in existing technologies are solved, achieving efficient image recognition and data security management.
Patent Information
- Authority / Receiving Office
- WO · WO
- Patent Type
- Applications
- Current Assignee / Owner
- CHI NING
- Filing Date
- 2026-01-15
- Publication Date
- 2026-07-23
Smart Images
Figure CN2026072819_23072026_PF_FP_ABST
Abstract
Description
Simplified Quantization Decomposition Logical Storage Architecture for Image Recognition Technical Field
[0001] This invention relates to the field of computer technology, specifically to a simplified quantization decomposition logic storage architecture for image recognition. Background Technology
[0002] In today's era of rapid digital information development, image recognition technologies are emerging in large numbers. However, existing technologies still have the following shortcomings when processing image, video, and other data: 1. When recognizing images, most existing technologies adopt a broad and coarse matching method, failing to fully utilize algorithms to simplify and classify images. Especially when facing massive amounts of image data, the lack of effective simplification strategies necessitates relying on high-frame-rate processors to maintain processing speed, which not only wastes hardware resources but also leads to low processing efficiency; 2. Existing image recognition technologies lack encryption measures, thus posing a serious risk of privacy leaks. Summary of the Invention
[0003] The purpose of this invention is to provide a simplified quantization decomposition logic storage architecture for image recognition, so as to solve the problems mentioned in the background art.
[0004] To achieve the above objectives, the present invention provides the following technical solution: a simplified image recognition quantization decomposition logic storage architecture, comprising the following steps: Step 1, image and video data collection; Step 2, feature value extraction; Step 3, combining data with letter sequence division; Step 4, establishment of letter sorting morphology quantization decomposition logic storage architecture; Step 5, matching of image and video data to be recognized; Step 6, recognition and determination.
[0005] In step one above, image and video data are collected and the data is preprocessed.
[0006] In step two above, feature values are extracted from the preprocessed image and video data.
[0007] In step three above, the first group uses the total number of images as the base and takes 1 / 24 of the total number of feature values as the initial retrieval standard. Data is filtered from the feature values, and any image containing a value that satisfies 1 / 24 of the total number of feature values is grouped into the first group. The first letter is labeled as A, and the specific conditions for feature value entry are written. The remaining 22 letters are labeled in the same way, and the last letter is a variable, set to not contain the above 23 numerical features. Each letter label in the first group represents 1 / 24 of the initial total number. The second group uses the 1 / 24 of the total number represented by the letters in the first group as the total number, and divides the data according to the division method of the first group of data, subdividing it 7 times to form an independent letter arrangement.
[0008] In step four above, all the images from step three are stored in folders, and documents are created inside the folders. Data features are written into cells, and data features are used for filtering. Each subdivision letter is used to label the document feature content and folder name. This process is repeated seven times, and the feature data of each subdivision is written into documents with multiple feature sets, thereby constructing a letter sorting form quantization decomposition logical storage architecture.
[0009] In step five above, feature values are extracted from the image and video data to be identified. The obtained feature values are then matched with the quantization decomposition logic storage architecture of the letter sorting pattern constructed in step four to obtain the letter sorting position of the image and video data to be identified, as well as the specific storage position data information.
[0010] In step six above, the similarity between the letter order of the image / video data to be identified and other images / video data is calculated. The identification is then determined based on the similarity calculation results, and the results are output.
[0011] Preferably, in step one, the image and video data preprocessing specifically includes: image data preprocessing includes image processing operations and numerical extraction operations; image processing operations include background removal, desaturation and inversion, and adjustment to line drawing; numerical extraction operations include key data extraction, grid division, and numerical annotation; and for video data, it is split into images and then processed in the same way as image preprocessing.
[0012] Preferably, the image processing operation is performed using Adobe Photoshop software, and the key data extraction is performed using AutoCAD software.
[0013] Preferably, in step three, when filtering data from feature values, single values that meet the feature value filtering conditions are selected first.
[0014] Preferably, in step six, the other image and video data are pre-stored in an alphabetical sorting morphological quantization decomposition logical storage architecture.
[0015] Preferably, in step six, the similarity calculation method adopts one of the following algorithms: edit distance algorithm based on character matching, cosine similarity algorithm based on letter frequency, and Jaccard similarity coefficient algorithm based on set theory.
[0016] Preferably, in step six, the identification and determination method is as follows: if the similarity reaches a set threshold, the data to be identified is determined to be successfully matched with the stored data, and the matched image or video data information is output; if the similarity does not reach the set threshold, no matching result is output.
[0017] Compared with the prior art, the beneficial effects of the present invention are as follows: The present invention generates independent logical storage space and independent identity letter sorting information by constructing a letter sorting morphology quantization decomposition logical storage architecture. This not only simplifies the calculation process and accelerates the address information transmission speed, enabling data processing to run efficiently under ordinary hardware configurations, but also greatly improves the recognition accuracy. Furthermore, the letter sorting morphology quantization decomposition logical storage architecture comes with encryption effects, which can effectively reduce the risk of privacy leakage in the distributed data storage recognition process, improve data security, and make it easier to manage. Attached Figure Description
[0018] Figure 1 is a flowchart of the method of the present invention;
[0019] Figure 2 is a schematic diagram of the data flow of the present invention. Detailed Implementation
[0020] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.
[0021] Please refer to Figures 1 and 2. One embodiment of the present invention provides a simplified image recognition quantization decomposition logic storage architecture, comprising the following steps: Step 1, collecting and modifying image and video data to a uniform specification; Step 2, feature extraction, extracting feature values of regions from images of the same specification using grid numerical annotation information, and setting threshold parameters; Step 3, combining the feature data contained in the threshold, and dividing using letter sequences; Step 4, establishing a letter sorting morphology quantization decomposition logic storage architecture; Step 5, matching the image and video data to be recognized; Step 6, recognition and determination.
[0022] In step one above, image and video data are collected and preprocessed. Specifically, image and video data preprocessing includes image processing and numerical extraction operations. Image processing operations include standardization, background removal, desaturation and inversion, and conversion to line art. Numerical extraction operations include key data extraction, grid generation, and numerical annotation. For video data, it is split into images and then processed using the same image preprocessing methods. Image processing is performed using Adobe Photoshop, and key data extraction is performed using AutoCAD.
[0023] In step two above, feature values are extracted from the preprocessed image and video data.
[0024] In step three above, 24 letters are grouped together. Taking the first group A as an example, the total number of images is used as the base, and 1 / 24 of the total number of feature values (the ideal value in aggregation theory) is taken as the initial retrieval standard. Data is filtered from the feature values, and a threshold is set. Images with feature values that meet the threshold and satisfy 1 / 24 of the total number of feature values are grouped into the first group, labeled with the first letter of the sequence A, and the specific conditions for feature value entry are written. The remaining 22 letters are labeled in the same way, and the 24th letter X is set as a variable to represent the features that do not contain the above 23 threshold numbers. Each letter label in the first group represents 1 / 24 of the initial total number. The second group uses the total number represented by the letters in the first group as the total number, and the data is divided according to the division method of the first group. The scheme takes 7 groups as an example. As shown in the case, the first four groups are the actual classification, and the last three groups are for completing independent identity information for a single image to achieve the decoupling requirements of the application. The number of groups should be based on the actual application requirements of this scheme. This includes reserving actual software expansion applications.
[0025] In step four above, all the images from step three are stored in folders, and documents are created inside the folders. Data features are written into cells, and threshold parameters are used to filter the data features. Each subdivision letter is used to label the document feature content and folder name. This process is repeated seven times, and the feature data of each subdivision is written into documents with multiple feature sets, thereby constructing a letter sorting morphological quantification decomposition logical storage architecture.
[0026] In step five above, feature values are extracted from the image and video data to be identified. The obtained feature values are then matched with the quantization decomposition logic storage architecture of the letter sorting pattern constructed in step four to obtain the letter sorting position of the image and video data to be identified, as well as the specific storage position data information.
[0027] In step six above, the letter order similarity between the image / video data to be identified and other image / video data is calculated. Recognition is then determined based on the similarity calculation results, and the results are output. The other image / video data is pre-stored in a letter order morphology quantization decomposition logical storage architecture. The similarity calculation method uses one of the following algorithms: edit distance algorithm based on character matching, cosine similarity algorithm based on letter frequency, or Jaccard similarity coefficient algorithm based on set theory. Specifically, if the similarity reaches a set threshold, the data to be identified is considered a successful match with the stored data, and the matched image / video data information is output. If the similarity does not reach the set threshold, no matching result is output.
[0028] Based on the above, the advantages of this invention are as follows: When used, this invention simplifies the image recognition and matching process by establishing a letter sorting morphological quantization decomposition logical storage architecture. Furthermore, it accelerates the transmission speed of address information by retrieving independent identity strings from dataset documents, reduces the requirements for device configuration, and enables data processing to run efficiently on ordinary hardware configurations. Moreover, the letter sorting morphological quantization decomposition logical storage architecture includes encryption effects, which are reflected in its logic. Specifically, for the grid, representative coordinate numbers can be freely set. Therefore, the planar numerical coordinate arrangement method, i.e., the letters represented by threshold features, has N decoupling methods. Features can be mapped to coordinates to find letters, and coordinates can be used to extract features to find algorithms and map to thresholds. Based on these multiple mapping decoupling relationships, the difficulty of external decryption is increased, effectively reducing the risk of privacy leakage and improving data security.
[0029] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the invention can be implemented in other specific forms without departing from its spirit or essential characteristics. Therefore, the embodiments should be considered in all respects as exemplary and non-limiting, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be included within the present invention. No reference numerals in the claims should be construed as limiting the scope of the claims.
Claims
1. A simplified image recognition quantization decomposition logical storage architecture, comprising the following steps: Step 1: Image and video data collection; Step 2: Feature value extraction; Step 3: Data segmentation using letter sequences; Step 4: Establishment of a logical storage structure for letter sorting morphology quantization decomposition; Step 5: Matching of images and videos to be identified; Step 6: Recognition and determination; Its key feature is: In step one above, image and video data are collected and the data is preprocessed. In step two above, feature values are extracted from the preprocessed image and video data. In step three above, the total number of images is used as the base, and 1 / 24 of the total number of feature values is taken as the initial retrieval standard. Data is filtered from the feature values, and any image containing a value that satisfies 1 / 24 of the total number of feature values is grouped into the first group. The first letter of the first group is labeled as A, and the specific conditions for feature value entry are written. The remaining 22 letters are labeled in the same way, and the last letter is a variable, which is set to not contain the above 23 numerical features. Each letter label in the first group represents 1 / 24 of the initial total number. The second group is divided according to the division method of the first group of data, and is subdivided into 7 times to form an independent letter arrangement. In step four above, all the images from step three are stored in folders, and documents are created inside the folders. Data features are written into cells, and data features are used for filtering. Each subdivision letter is used to label the document feature content and folder name. This process is repeated seven times, and the feature data of each subdivision is written into documents with multiple feature sets, thereby constructing a letter sorting form quantization decomposition logical storage architecture. In step five above, feature values are extracted from the image and video data to be identified. The obtained feature values are then matched with the quantization decomposition logic storage architecture of the letter sorting pattern constructed in step four to obtain the letter sorting position of the image and video data to be identified, as well as the specific storage position data information. In step six above, the similarity between the letter order of the image / video data to be identified and other images / video data is calculated. The identification is then determined based on the similarity calculation results, and the results are output.
2. The simplified image recognition quantization decomposition logic storage architecture according to claim 1, characterized in that: In step one, the image and video data preprocessing specifically includes: image data preprocessing includes image processing operations and numerical extraction operations. Image processing operations include background removal, desaturation and inversion, and adjustment to line drawing. Numerical extraction operations include key data extraction, grid division, and numerical annotation. For video data, it is split into images and then processed in the same way as image preprocessing.
3. The simplified image recognition quantization decomposition logic storage architecture according to claim 2, characterized in that: The image processing operations were performed using Adobe Photoshop, and the key data extraction was performed using AutoCAD.
4. The simplified image recognition quantization decomposition logic storage architecture according to claim 1, characterized in that: In step three, when filtering data from feature values, priority is given to selecting single values that meet the feature value filtering conditions.
5. The simplified image recognition quantization decomposition logic storage architecture according to claim 1, characterized in that: In step six, the other image and video data are pre-stored in the alphabetical sorting morphological quantization decomposition logical storage architecture.
6. The simplified image recognition quantization decomposition logic storage architecture according to claim 1, characterized in that: In step six, the similarity calculation method adopts one of the following algorithms: edit distance algorithm based on character matching, cosine similarity algorithm based on letter frequency, and Jaccard similarity coefficient algorithm based on set theory.
7. The simplified image recognition quantization decomposition logic storage architecture according to claim 1, characterized in that: In step six, the identification and judgment method is as follows: if the similarity reaches the set threshold, it is determined that the data to be identified is successfully matched with the stored data, and the matched image or video data information is output; if the similarity does not reach the set threshold, no matching result is output.