CN115205952B - Online learning image acquisition method and system based on deep learning - Google Patents
Online learning image acquisition method and system based on deep learning Download PDFInfo
- Publication number
- CN115205952B CN115205952B CN202211128260.8A CN202211128260A CN115205952B CN 115205952 B CN115205952 B CN 115205952B CN 202211128260 A CN202211128260 A CN 202211128260A CN 115205952 B CN115205952 B CN 115205952B
- Authority
- CN
- China
- Prior art keywords
- data
- image
- image data
- portrait
- characteristic
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/16—Human faces, e.g. facial parts, sketches or expressions
- G06V40/168—Feature extraction; Face representation
- G06V40/171—Local features and components; Facial parts ; Occluding parts, e.g. glasses; Geometrical relationships
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/10—Image acquisition
- G06V10/12—Details of acquisition arrangements; Constructional details thereof
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/26—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/30—Noise filtering
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/34—Smoothing or thinning of the pattern; Morphological operations; Skeletonisation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/44—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/74—Image or video pattern matching; Proximity measures in feature spaces
- G06V10/761—Proximity, similarity or dissimilarity measures
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/764—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using classification, e.g. of video objects
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/82—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using neural networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/16—Human faces, e.g. facial parts, sketches or expressions
- G06V40/172—Classification, e.g. identification
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Computation (AREA)
- Computing Systems (AREA)
- Artificial Intelligence (AREA)
- Oral & Maxillofacial Surgery (AREA)
- Software Systems (AREA)
- Databases & Information Systems (AREA)
- Medical Informatics (AREA)
- Human Computer Interaction (AREA)
- Life Sciences & Earth Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Molecular Biology (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses an online learning image acquisition method and system based on deep learning. The method comprises the steps of obtaining image data abnormal information through obtaining initial image data of a target object, performing image preprocessing and data abnormal analysis, performing targeted data abnormal recovery according to the data abnormal information, and obtaining image data after abnormal processing as transmission data, so that the data quality after image data acquisition is greatly improved, the error occurrence situation of data transmission caused by data abnormality is reduced, and the acquisition efficiency of high-quality data is improved. In addition, the invention also transmits the image data through image matrixing encryption, thereby improving the safety and the transmission efficiency of the image data.
Description
Technical Field
The invention relates to the field of deep learning, in particular to an online learning image acquisition method and system based on deep learning.
Background
At present, with the gradual development of information technology and the great expansion of online education requirements, the trend that the traditional learning is changed into online learning on the internet is more and more obvious, and the development of online learning becomes the current hot topic.
However, when learning to collect user image data on line, the situation that the collected image data has inconsistent formats is often caused due to the difference of the user terminal camera devices, and when learning to collect image data on line, due to the abnormal situations that data is often lost, wrong, image information is incomplete and the like in the transmission process of the data, if the image data is directly transmitted, the transmission of the image data is easy to cause errors. Therefore, a method for improving the quality of the acquired image data is needed to solve the above problems.
Disclosure of Invention
In order to solve at least one technical problem, the invention provides an online learning image acquisition based on deep learning.
The invention provides an online learning image acquisition method based on deep learning, which comprises the following steps:
performing online learning image sampling according to a preset sampling rate to obtain initial image data of a target object;
carrying out data preprocessing on the initial image data according to a preset transmission standard to obtain transmission image data;
carrying out data anomaly analysis and classification on the transmitted image data to obtain data anomaly classification information;
performing image processing and data updating recovery based on a convolutional neural network according to the data anomaly classification information and the transmission image data to obtain updated transmission image data;
and encrypting, compressing and transmitting the updated and transmitted image data to a preset server.
In this scheme, the sampling of the online learning image is performed according to a preset sampling rate to obtain initial image data of the target object, and the method includes the following steps:
according to a preset sampling rate, carrying out image sampling from a user terminal to obtain target object reference image data;
performing data anomaly analysis and screening on the target object reference image data to obtain non-anomalous image data;
carrying out image characteristic analysis on the non-abnormal image data, screening out image data with the image characteristic degree larger than a preset characteristic degree, and obtaining high-characteristic contrast image data;
and constructing a history contrast image database, and importing the high-feature contrast image data into the history contrast image database.
In this scheme, the data preprocessing is performed on the initial image data according to a preset transmission standard to obtain transmission image data, and specifically includes:
acquiring initial image data, and judging the storage format of the initial image data;
if the initial image data storage format is not the preset image storage format, performing lossless conversion on the preset image storage format to obtain converted image data;
and performing image data geometric transformation processing on the converted image data according to the preset image size and resolution to obtain transmission image data.
In this scheme, the data anomaly analysis and classification of the transmission image data are performed to obtain data anomaly classification information, which specifically includes:
carrying out image smoothing and noise reduction preprocessing on the transmission image data to obtain enhanced image data;
carrying out portrait area identification and image segmentation on the enhanced image data to obtain portrait area image data and background area image data;
constructing an image feature recognition model based on a single-layer convolutional neural network;
importing portrait region image data and background region image data into the image feature identification model for image feature extraction to obtain portrait feature data and background feature data;
performing image characteristic anomaly analysis on the portrait characteristic data and the background characteristic data to obtain portrait data anomaly information and background data anomaly information; and merging the portrait data abnormal information and the background data abnormal information to obtain data abnormal classification information.
In this scheme, the image processing and data update recovery based on the convolutional neural network are performed according to the data anomaly classification information and the transmission image data to obtain updated transmission image data, specifically:
acquiring portrait data abnormal information in the data abnormal classification information;
acquiring a target object retrieval tag, and retrieving from a historical comparison image database according to the target object retrieval tag to obtain target object comparison image data;
carrying out portrait feature analysis on the contrast image data to obtain contrast portrait features;
and performing data exception recovery processing on the portrait feature data according to the portrait data exception information and the comparative portrait features to obtain second portrait feature data.
In this scheme, the image processing and data update recovery based on the convolutional neural network are performed according to the data anomaly classification information and the transmission image data to obtain updated transmission image data, further including:
acquiring background data abnormal information in the data abnormal classification information;
acquiring background characteristic data, and performing characteristic retrieval on the background characteristic data from image big data to obtain retrieval characteristic data higher than a preset similarity;
according to the background data abnormal information, in combination with the retrieval characteristic data, performing data abnormal recovery processing on the background characteristic data to obtain second background characteristic data;
and carrying out image anomaly correction on the transmission image data according to the second portrait characteristic data and the second background characteristic data to obtain updated transmission image data.
In this scheme, the encrypting, compressing and transmitting the updated and transmitted image data to the preset server specifically includes:
performing matrix conversion on the updated and transmitted image data to obtain an image matrix;
dividing data according to the size of the image matrix to obtain N sub-matrixes;
carrying out data encryption on the data corresponding to the N sub-matrixes according to a preset encryption algorithm to obtain N encrypted data blocks;
merging and compressing the N encrypted data blocks to obtain result transmission data;
and sending the result transmission data to a preset server for data storage.
The second aspect of the present invention also provides an online learning image acquisition system based on deep learning, including: the on-line learning image acquisition method based on the deep learning is implemented by the processor, and the following steps are implemented when the on-line learning image acquisition method based on the deep learning is executed by the processor:
performing online learning image sampling according to a preset sampling rate to obtain initial image data of a target object;
carrying out data preprocessing on the initial image data according to a preset transmission standard to obtain transmission image data;
carrying out data anomaly analysis and classification on the transmitted image data to obtain data anomaly classification information;
performing image processing and data updating recovery based on a convolutional neural network according to the data anomaly classification information and the transmission image data to obtain updated transmission image data;
and encrypting, compressing and transmitting the updated transmission image data to a preset server.
In this scheme, the sampling of the online learning image is performed according to a preset sampling rate to obtain initial image data of the target object, and the method includes the following steps:
according to a preset sampling rate, carrying out image sampling from a user terminal to obtain target object reference image data;
performing data anomaly analysis and screening on the target object reference image data to obtain non-anomalous image data;
carrying out image characteristic analysis on the non-abnormal image data, screening out image data with the image characteristic degree larger than a preset characteristic degree, and obtaining high-characteristic contrast image data;
and constructing a history contrast image database, and importing the high-feature contrast image data into the history contrast image database.
In this scheme, the encrypting, compressing and transmitting the updated and transmitted image data to the preset server specifically includes:
performing matrix conversion on the updated and transmitted image data to obtain an image matrix;
dividing data according to the size of the image matrix to obtain N sub-matrixes;
carrying out data encryption on the data corresponding to the N sub-matrixes according to a preset encryption algorithm to obtain N encrypted data blocks;
merging and compressing the N encrypted data blocks to obtain result transmission data;
and sending the result transmission data to a preset server for data storage.
The third aspect of the present invention further provides a computer-readable storage medium, where the computer-readable storage medium includes a program of an on-line learning image acquisition method based on deep learning, and when the program of the on-line learning image acquisition method based on deep learning is executed by a processor, the step of the on-line learning image acquisition method based on deep learning is implemented as any one of the above steps.
The invention discloses an online learning image acquisition method and system based on deep learning. The method comprises the steps of obtaining image data abnormal information through obtaining initial image data of a target object, performing image preprocessing and data abnormal analysis, performing targeted data abnormal recovery according to the data abnormal information, and obtaining image data after abnormal processing as transmission data, so that the data quality after image data acquisition is greatly improved, the error occurrence situation of data transmission caused by data abnormality is reduced, and the acquisition efficiency of high-quality data is improved. In addition, the invention also transmits the image data through image matrixing encryption, thereby improving the safety and the transmission efficiency of the image data.
Drawings
FIG. 1 is a flow chart of an online learning image acquisition method based on deep learning according to the present invention;
FIG. 2 is a flow chart illustrating the process of acquiring transmission image data according to the present invention;
FIG. 3 is a flow chart illustrating the process of obtaining second portrait session data according to the present invention;
fig. 4 shows a block diagram of an online learning image acquisition system based on deep learning according to the present invention.
Detailed Description
In order that the above objects, features and advantages of the present invention can be more clearly understood, a more particular description of the invention will be rendered by reference to the appended drawings. It should be noted that the embodiments and features of the embodiments of the present application may be combined with each other without conflict.
In the following description, numerous specific details are set forth in order to provide a thorough understanding of the present invention, however, the present invention may be practiced otherwise than as specifically described herein and, therefore, the scope of the present invention is not limited by the specific embodiments disclosed below.
Fig. 1 shows a flowchart of an online learning image acquisition method based on deep learning according to the present invention.
As shown in fig. 1, a first aspect of the present invention provides an online learning image acquisition method based on deep learning, including:
s102, sampling an online learning image according to a preset sampling rate to obtain initial image data of a target object;
s104, performing data preprocessing on the initial image data according to a preset transmission standard to obtain transmission image data;
s106, performing data anomaly analysis and classification on the transmission image data to obtain data anomaly classification information;
s108, performing image processing and data updating recovery based on the convolutional neural network according to the data abnormal classification information and the transmission image data to obtain updated transmission image data;
and S110, encrypting and compressing the updated transmission image data and transmitting the encrypted transmission image data to a preset server.
According to the embodiment of the present invention, the on-line learning image sampling according to the preset sampling rate to obtain the initial image data of the target object includes:
according to a preset sampling rate, carrying out image sampling from a user terminal to obtain target object reference image data;
performing data anomaly analysis and screening on the target object reference image data to obtain non-anomalous image data;
carrying out image characteristic analysis on the non-abnormal image data, screening out image data with the image characteristic degree larger than a preset characteristic degree, and obtaining high-characteristic contrast image data;
and constructing a history contrast image database, and importing the high-feature contrast image data into the history contrast image database.
It should be noted that the preset sampling rate is generally a general sampling rate for network image transmission, and the data anomaly analysis and screening includes determining abnormal conditions such as data loss, data partial loss, data error and the like. And in the screened image data with the image characteristic degree larger than the preset characteristic degree, the image characteristic degree is a main index reflecting the image distinguishing characteristic.
According to the embodiment of the invention, the method further comprises the following steps:
acquiring parameter information of user terminal camera equipment;
generating an acquired image parameter according to the parameter information;
generating user image acquisition standard information according to the acquired image parameters;
acquiring target image acquisition standard information, and generating an image standard conversion model by combining with user image acquisition standard information;
and importing the initial image data of the target object into an image standard conversion model for image format standardization conversion.
It should be noted that the user image acquisition standard information includes information such as the size, resolution, sampling rate, and color saturation of an image during user image acquisition.
According to the embodiment of the invention, the method further comprises the following steps:
acquiring initial image data of a target object;
carrying out gray level conversion on the initial image data to obtain gray level image data;
performing matrix conversion on the gray image data to obtain a gray image matrix;
dividing the gray image matrix longitudinally and transversely to obtain M1 multiplied by M2 target sub-matrixes;
retrieving corresponding high-feature contrast image data from a history contrast image database according to the target object;
carrying out gray level conversion and matrix segmentation on the high-feature contrast image data to obtain M1 × M2 contrast submatrices;
performing matrix data comparison analysis on the target sub-matrix and the comparison sub-matrix, and calculating the data similarity of each target sub-matrix and the comparison sub-matrix;
and if the M1 multiplied by M2 data similarity degrees are all larger than the preset matrix similarity degree, judging that the initial image data of the target object is historical repeated data, and sending the repeated data information to the user terminal for secondary data acquisition.
It should be noted that after the initial image data of the target object is acquired, since uncertain factors such as abnormality may occur in the user terminal image pickup device, the acquired image data is a frame where the last image pickup stays, and at this time, the acquired image data has high coincidence with the historical data, and it is necessary to make a corresponding judgment analysis and acquire the secondary data, so as to obtain accurate initial image data of the target object. In addition, the repeated data can be efficiently judged through the similarity analysis and judgment of the image matrix, so that the efficiency of effective data acquisition is improved. The size of M1 and M2 is determined by the data size of the grayscale image data, and the larger the data size, the larger the values of M1 and M2.
FIG. 2 shows a flow chart for acquiring transmission image data according to the present invention.
According to the embodiment of the present invention, the data preprocessing is performed on the initial image data according to a preset transmission standard to obtain the transmission image data, and specifically, the data preprocessing includes:
s202, acquiring initial image data and judging the storage format of the initial image data;
s204, if the initial image data storage format is not the preset image storage format, carrying out lossless conversion on the preset image storage format to obtain converted image data;
and S206, carrying out image data geometric transformation processing on the converted image data according to the preset image size and resolution to obtain transmission image data.
It should be noted that the preset transmission standard includes a preset image storage format, a preset image size and a preset image resolution, and the preset image storage format includes BMP, JPG, and PBG.
According to the embodiment of the present invention, the data anomaly analysis and classification of the transmission image data to obtain data anomaly classification information specifically includes:
carrying out image smoothing and noise reduction preprocessing on the transmission image data to obtain enhanced image data;
carrying out portrait area identification and image segmentation on the enhanced image data to obtain portrait area image data and background area image data;
constructing an image feature recognition model based on a single-layer convolutional neural network;
importing portrait region image data and background region image data into the image feature identification model for image feature extraction to obtain portrait feature data and background feature data;
performing image characteristic anomaly analysis on the portrait characteristic data and the background characteristic data to obtain portrait data anomaly information and background data anomaly information; and merging the portrait data abnormal information and the background data abnormal information to obtain data abnormal classification information.
It should be noted that the image characteristic abnormality includes singularization of a characteristic data pixel, loss of characteristic data, color saturation and gray level abnormality of characteristic data, and the like. The image feature recognition model based on the single-layer convolutional neural network is a deep learning network model.
In addition, the data anomaly classification information further comprises a characteristic anomaly index, and the specific calculation formula of the characteristic anomaly index is as follows:
wherein the content of the first and second substances,
in order to be a characteristic abnormality index,
is the gray value of the ith pixel, M is the total number of the image pixels,
is the average degree of the colors of the image,
is the average gray level of the image, T is the number of color channels,
is the color value of the jth channel.
Fig. 3 shows a flow chart of the invention for obtaining second portrait session data.
According to the embodiment of the present invention, the image processing and data update recovery based on the convolutional neural network are performed according to the data anomaly classification information and the transmission image data to obtain the updated transmission image data, specifically:
s302, obtaining portrait data abnormal information in the data abnormal classification information;
s304, acquiring a target object retrieval tag, and retrieving from a historical comparison image database according to the target object retrieval tag to obtain target object comparison image data;
s306, performing portrait feature analysis on the contrast image data to obtain contrast portrait features;
and S308, performing data exception recovery processing on the portrait feature data according to the portrait data exception information and the comparative portrait features to obtain second portrait feature data.
According to the embodiment of the present invention, the performing image processing and data update recovery based on a convolutional neural network according to the data anomaly classification information and the transmission image data to obtain updated transmission image data further includes:
acquiring background data abnormal information in the data abnormal classification information;
acquiring background characteristic data, and performing characteristic retrieval on the background characteristic data from image big data to obtain retrieval characteristic data higher than a preset similarity;
according to the background data abnormal information, in combination with the retrieval characteristic data, performing data abnormal recovery processing on the background characteristic data to obtain second background characteristic data;
and according to the second portrait characteristic data and the second background characteristic data, carrying out image abnormity correction on the transmission image data to obtain updated transmission image data.
It should be noted that, compared with the transmission image data, the updating transmission image data repairs the abnormal data part and improves the data quality of the transmission image, so as to more efficiently acquire and transmit the image data.
It is worth mentioning that in the process of acquiring image data, situations such as image distortion and pixel blurring often occur, and by analyzing the abnormal situation of the image data and performing targeted abnormal recovery of the image data according to the abnormal situation, normal image data meeting the transmission data standard can be obtained.
According to the embodiment of the present invention, the image anomaly correction of the transmission image data further includes:
acquiring portrait feature data of a target object;
performing facial feature analysis according to the portrait feature data to obtain facial region information;
carrying out proportion analysis according to the region information of the facial features to obtain specific proportion information of the facial features;
carrying out attitude characteristic analysis according to the portrait characteristic data to obtain specific attitude proportion information;
carrying out information data integration on the specific facial feature proportion information and the specific posture proportion information to obtain target object identification proportion data;
importing the target object recognition ratio data into a history contrast image database for storage;
and carrying out image anomaly correction on the transmission image data according to the target object identification ratio data.
It should be noted that the specific facial feature proportion information is relative proportion information of positions and sizes of facial features of the target object, and the specific posture proportion information is proportion information between the region images represented by the posture contour of the target object. By analyzing the target object recognition ratio data, the efficiency of image abnormality correction can be improved.
According to the embodiment of the invention, the method further comprises the following steps:
carrying out contour feature analysis on the updated and transmitted image data to obtain portrait contour feature data;
performing portrait posture analysis according to the portrait contour characteristic data to obtain current posture characteristic data;
performing attitude comparison analysis according to the current attitude characteristic data and preset attitude characteristic data to obtain an attitude similarity index;
carrying out attitude characteristic contour consistency analysis on the current attitude characteristic data to obtain a contour consistency coefficient;
performing portrait background characteristic analysis on the updated and transmitted image data to obtain portrait area proportion information and background area proportion information;
calculating to obtain an on-line learning image quality evaluation coefficient according to the attitude similarity index, the contour consistency coefficient, the portrait area proportion information and the background area proportion information;
and if the on-line learning image quality evaluation coefficient is smaller than the preset evaluation coefficient, performing secondary data updating recovery on the updated and transmitted image data.
The specific calculation formula of the on-line learning image quality evaluation coefficient is as follows:
wherein, the first and the second end of the pipe are connected with each other,
in order to learn the image quality evaluation coefficient on line,
the ratio of the area of the portrait is,
in order to make the background region a high ratio,
in order to be an index of the attitude similarity,
is the contour continuity coefficient.
In the updating of the transmitted image data, after the data is abnormally restored, if the degree of distortion of the posture feature of the target object in the image data is higher, the corresponding contour continuity coefficient is lower. The preset posture characteristic data is generally sitting posture characteristic data.
According to the embodiment of the present invention, the encrypting, compressing and transmitting the updated and transmitted image data to the preset server specifically includes:
performing matrix conversion on the updated and transmitted image data to obtain an image matrix;
dividing data according to the size of the image matrix to obtain N sub-matrixes;
carrying out data encryption on the data corresponding to the N sub-matrixes according to a preset encryption algorithm to obtain N encrypted data blocks;
merging and compressing the N encrypted data blocks to obtain result transmission data;
and sending the result transmission data to a preset server for data storage.
It should be noted that the preset encryption algorithm includes an AES encryption algorithm, an RSA encryption algorithm, an MD5 encryption algorithm, and a DES encryption algorithm. In addition, matrix conversion and data division encryption transmission are carried out on the updated and transmitted image data, and therefore the safety and the transmission efficiency of the image data can be improved. And performing data division according to the size of the image matrix to obtain N sub-matrixes, wherein the larger the data volume of the updated and transmitted image data is, the larger the size of the image matrix is, and the larger the corresponding data volume is.
Fig. 4 shows a block diagram of an online learning image acquisition system based on deep learning according to the present invention.
The second aspect of the present invention also provides an online learning image acquisition system 4 based on deep learning, which includes: a memory 41 and a processor 42, wherein the memory includes a program of an on-line learning image acquisition method based on deep learning, and when the program of the on-line learning image acquisition method based on deep learning is executed by the processor, the following steps are implemented:
performing online learning image sampling according to a preset sampling rate to obtain initial image data of a target object;
carrying out data preprocessing on the initial image data according to a preset transmission standard to obtain transmission image data;
carrying out data anomaly analysis and classification on the transmitted image data to obtain data anomaly classification information;
performing image processing and data updating recovery based on a convolutional neural network according to the data anomaly classification information and the transmission image data to obtain updated transmission image data;
and encrypting, compressing and transmitting the updated and transmitted image data to a preset server.
According to the embodiment of the present invention, the on-line learning image sampling according to the preset sampling rate to obtain the initial image data of the target object includes:
according to a preset sampling rate, carrying out image sampling from a user terminal to obtain target object reference image data;
performing data anomaly analysis and screening on the target object reference image data to obtain non-anomalous image data;
carrying out image characteristic analysis on the non-abnormal image data, screening out image data with the image characteristic degree larger than a preset characteristic degree, and obtaining high-characteristic contrast image data;
and constructing a history contrast image database, and importing the high-feature contrast image data into the history contrast image database.
It should be noted that the preset sampling rate is generally a general sampling rate for network image transmission, and the data anomaly analysis and screening includes determining abnormal conditions such as data loss, data partial loss, data error and the like. And in the screened image data with the image characteristic degree larger than the preset characteristic degree, the image characteristic degree is a main index reflecting the image distinguishing characteristic.
According to the embodiment of the invention, the method further comprises the following steps:
acquiring parameter information of user terminal camera equipment;
generating an acquired image parameter according to the parameter information;
generating user image acquisition standard information according to the acquired image parameters;
acquiring target image acquisition standard information, and generating an image standard conversion model by combining the user image acquisition standard information;
and importing the initial image data of the target object into an image standard conversion model for image format standardization conversion.
It should be noted that the user image acquisition standard information includes information such as the size, resolution, sampling rate, and color saturation of an image during user image acquisition.
According to the embodiment of the invention, the method further comprises the following steps:
acquiring initial image data of a target object;
carrying out gray level conversion on the initial image data to obtain gray level image data;
performing matrix conversion on the gray image data to obtain a gray image matrix;
dividing the gray image matrix longitudinally and transversely to obtain M1 multiplied by M2 target sub-matrixes;
retrieving corresponding high-feature contrast image data from a history contrast image database according to the target object;
carrying out gray level conversion and matrix segmentation on the high-feature contrast image data to obtain M1 multiplied by M2 contrast submatrices;
performing matrix data comparison analysis on the target sub-matrix and the contrast sub-matrix, and calculating the data similarity of each target sub-matrix and each contrast sub-matrix;
and if the M1 multiplied by M2 data similarity degrees are all larger than the preset matrix similarity degree, judging that the initial image data of the target object is historical repeated data, and sending the repeated data information to the user terminal for secondary data acquisition.
It should be noted that after the initial image data of the target object is acquired, since uncertain factors such as abnormality may occur in the user terminal image pickup device, the acquired image data is a frame where the last image pickup stays, and at this time, the acquired image data has high coincidence with the historical data, and it is necessary to make a corresponding judgment analysis and acquire the secondary data, so as to obtain accurate initial image data of the target object. In addition, the repeated data can be efficiently judged through the similarity analysis and judgment of the image matrix, so that the efficiency of effective data acquisition is improved. The size of M1 and M2 is determined by the data size of the grayscale image data, and the larger the data size, the larger the values of M1 and M2.
According to the embodiment of the present invention, the data preprocessing is performed on the initial image data according to a preset transmission standard to obtain the transmission image data, and specifically, the data preprocessing includes:
acquiring initial image data and judging the storage format of the initial image data;
if the initial image data storage format is not the preset image storage format, performing lossless conversion on the preset image storage format to obtain converted image data;
and performing image data geometric transformation processing on the converted image data according to the preset image size and resolution to obtain transmission image data.
It should be noted that the preset transmission standard includes a preset image storage format, a preset image size and a preset image resolution, and the preset image storage format includes BMP, JPG, and PBG.
According to the embodiment of the present invention, the data anomaly analysis and classification of the transmission image data to obtain data anomaly classification information specifically includes:
carrying out image smoothing and noise reduction preprocessing on the transmission image data to obtain enhanced image data;
carrying out portrait area identification and image segmentation on the enhanced image data to obtain portrait area image data and background area image data;
constructing an image feature recognition model based on a single-layer convolution neural network;
importing portrait region image data and background region image data into the image feature identification model for image feature extraction to obtain portrait feature data and background feature data;
performing image characteristic anomaly analysis on the portrait characteristic data and the background characteristic data to obtain portrait data anomaly information and background data anomaly information; and merging the portrait data abnormal information and the background data abnormal information to obtain data abnormal classification information.
It should be noted that the image characteristic abnormality includes singularization of characteristic data pixel points, characteristic data loss, characteristic data chroma and gray level abnormality, and the like. The image feature recognition model based on the single-layer convolutional neural network is a deep learning network model.
In addition, the data abnormality classification information further includes a characteristic abnormality index, and the specific calculation formula of the characteristic abnormality index is as follows:
wherein, the first and the second end of the pipe are connected with each other,
in order to be a characteristic abnormality index,
is the gray value of the ith pixel, M is the total number of the image pixels,
is the average degree of the colors of the image,
is the average gray level of the image, T is the number of color channels,
is the color value of the jth channel.
According to the embodiment of the present invention, the image processing and data update recovery based on the convolutional neural network are performed according to the data anomaly classification information and the transmission image data to obtain the updated transmission image data, specifically:
acquiring portrait data abnormal information in the data abnormal classification information;
acquiring a target object retrieval tag, and retrieving from a history comparison image database according to the target object retrieval tag to obtain target object comparison image data;
carrying out portrait feature analysis on the contrast image data to obtain contrast portrait features;
and performing data exception recovery processing on the portrait feature data according to the portrait data exception information and the comparative portrait features to obtain second portrait feature data.
According to the embodiment of the present invention, the image processing and data update recovery based on a convolutional neural network are performed according to the data anomaly classification information and the transmission image data to obtain updated transmission image data, and the method further includes:
acquiring background data abnormal information in the data abnormal classification information;
acquiring background characteristic data, and performing characteristic retrieval on the background characteristic data from image big data to obtain retrieval characteristic data higher than a preset similarity;
according to the background data abnormal information, in combination with the retrieval characteristic data, performing data abnormal recovery processing on the background characteristic data to obtain second background characteristic data;
and according to the second portrait characteristic data and the second background characteristic data, carrying out image abnormity correction on the transmission image data to obtain updated transmission image data.
It should be noted that, compared with the transmission image data, the updating transmission image data repairs the abnormal part of the data, and improves the data quality of the transmission image, so as to more efficiently acquire and transmit the image data.
It should be noted that in the process of acquiring image data, conditions such as image distortion and pixel blurring often occur, and by analyzing abnormal conditions of the image data, targeted abnormal recovery of the image data is performed according to the abnormal conditions, so that normal image data meeting the transmission data standard can be obtained.
According to the embodiment of the present invention, the image anomaly correction of the transmission image data further includes:
acquiring portrait feature data of a target object;
performing facial feature analysis according to the portrait feature data to obtain facial region information;
carrying out proportion analysis according to the region information of the facial features to obtain specific proportion information of the facial features;
carrying out attitude characteristic analysis according to the portrait characteristic data to obtain specific attitude proportion information;
carrying out information data integration on the specific facial feature proportion information and the specific posture proportion information to obtain target object identification proportion data;
importing the target object recognition ratio data into a history contrast image database for storage;
and carrying out image anomaly correction on the transmission image data according to the target object identification ratio data.
It should be noted that the specific facial feature proportion information is relative proportion information of positions and sizes of facial features of the target object, and the specific posture proportion information is proportion information between the region images represented by the posture contour of the target object. By analyzing the target object recognition proportion data, the efficiency of image anomaly correction can be improved.
According to the embodiment of the invention, the method further comprises the following steps:
carrying out contour feature analysis on the updated and transmitted image data to obtain portrait contour feature data;
performing portrait attitude analysis according to the portrait contour characteristic data to obtain current attitude characteristic data;
performing attitude comparison analysis according to the current attitude characteristic data and preset attitude characteristic data to obtain an attitude similarity index;
carrying out attitude characteristic contour consistency analysis on the current attitude characteristic data to obtain a contour consistency coefficient;
performing portrait background characteristic analysis on the updated and transmitted image data to obtain portrait area proportion information and background area proportion information;
calculating to obtain an on-line learning image quality evaluation coefficient according to the attitude similarity index, the contour consistency coefficient, the portrait area proportion information and the background area proportion information;
and if the on-line learning image quality evaluation coefficient is smaller than the preset evaluation coefficient, performing secondary data updating recovery on the updated and transmitted image data.
The specific calculation formula of the on-line learning image quality evaluation coefficient is as follows:
wherein the content of the first and second substances,
in order to learn the image quality evaluation coefficient on line,
the ratio of the area occupied by the portrait is,
in order to make the background region a high ratio,
in order to be an index of the attitude similarity,
is the contour continuity coefficient.
In the updating of the transmitted image data, after the data is abnormally restored, if the degree of distortion of the posture feature of the target object in the image data is higher, the corresponding contour continuity coefficient is lower. The preset posture characteristic data is generally sitting posture characteristic data.
According to the embodiment of the present invention, the encrypting, compressing and transmitting the updated and transmitted image data to the preset server specifically includes:
carrying out matrix conversion on the updated and transmitted image data to obtain an image matrix;
dividing data according to the size of the image matrix to obtain N sub-matrixes;
carrying out data encryption on the data corresponding to the N sub-matrixes according to a preset encryption algorithm to obtain N encrypted data blocks;
merging and compressing the N encrypted data blocks to obtain result transmission data;
and sending the result transmission data to a preset server for data storage.
It should be noted that the preset encryption algorithm includes an AES encryption algorithm, an RSA encryption algorithm, an MD5 encryption algorithm, and a DES encryption algorithm. In addition, matrix conversion and data division encryption transmission are carried out on the updated and transmitted image data, and therefore the safety and the transmission efficiency of the image data can be improved. And performing data division according to the size of the image matrix to obtain N sub-matrixes, wherein the larger the data volume of the updated and transmitted image data is, the larger the size of the image matrix is, and the larger the corresponding data volume is.
The third aspect of the present invention further provides a computer-readable storage medium, where the computer-readable storage medium includes a program of an on-line learning image acquisition method based on deep learning, and when the program of the on-line learning image acquisition method based on deep learning is executed by a processor, the method implements the steps of the on-line learning image acquisition method based on deep learning according to any one of the above items.
The invention discloses an online learning image acquisition method and system based on deep learning. The method comprises the steps of obtaining image data abnormal information through obtaining initial image data of a target object, performing image preprocessing and data abnormal analysis, performing targeted data abnormal recovery according to the data abnormal information, and obtaining image data after abnormal processing as transmission data, so that the data quality after image data acquisition is greatly improved, the error occurrence situation of data transmission caused by data abnormality is reduced, and the acquisition efficiency of high-quality data is improved. In addition, the invention also transmits the image data through image matrixing encryption, thereby improving the safety and the transmission efficiency of the image data.
In the several embodiments provided in the present application, it should be understood that the disclosed apparatus and method may be implemented in other manners. The above-described device embodiments are merely illustrative, for example, the division of the unit is only a logical functional division, and there may be other division ways in actual implementation, such as: multiple units or components may be combined, or may be integrated into another system, or some features may be omitted, or not implemented. In addition, the coupling, direct coupling or communication connection between the components shown or discussed may be through some interfaces, and the indirect coupling or communication connection between the devices or units may be electrical, mechanical or in other forms.
The units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units; can be located in one place or distributed on a plurality of network units; some or all of the units can be selected according to actual needs to achieve the purpose of the solution of the embodiment.
In addition, all the functional units in the embodiments of the present invention may be integrated into one processing unit, or each unit may be separately regarded as one unit, or two or more units may be integrated into one unit; the integrated unit can be realized in a form of hardware, or in a form of hardware plus a software functional unit.
Those of ordinary skill in the art will understand that: all or part of the steps for realizing the method embodiments can be completed by hardware related to program instructions, the program can be stored in a computer readable storage medium, and the program executes the steps comprising the method embodiments when executed; and the aforementioned storage medium includes: a mobile storage device, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk or an optical disk, and other various media capable of storing program codes.
Alternatively, the integrated unit of the present invention may be stored in a computer-readable storage medium if it is implemented in the form of a software functional module and sold or used as a separate product. Based on such understanding, the technical solutions of the embodiments of the present invention or portions thereof contributing to the prior art may be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for enabling a computer device (which may be a personal computer, a server, or a network device) to execute all or part of the methods described in the embodiments of the present invention. And the aforementioned storage medium includes: a removable storage device, a ROM, a RAM, a magnetic or optical disk, or various other media capable of storing program code.
The above description is only for the specific embodiments of the present invention, but the scope of the present invention is not limited thereto, and any person skilled in the art can easily conceive of the changes or substitutions within the technical scope of the present invention, and all the changes or substitutions should be covered within the scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the appended claims.
Claims (7)
1. An online learning image acquisition method based on deep learning is characterized by comprising the following steps:
performing online learning image sampling according to a preset sampling rate to obtain initial image data of a target object;
carrying out data preprocessing on the initial image data according to a preset transmission standard to obtain transmission image data;
carrying out data anomaly analysis and classification on the transmitted image data to obtain data anomaly classification information;
performing image processing and data updating recovery based on a convolutional neural network according to the data anomaly classification information and the transmission image data to obtain updated transmission image data;
encrypting, compressing and transmitting the updated and transmitted image data to a preset server;
the method comprises the following steps of performing data anomaly analysis and classification on transmission image data to obtain data anomaly classification information, and specifically comprises the following steps:
carrying out image smoothing and noise reduction pretreatment on the transmission image data to obtain enhanced image data;
performing portrait area identification and image segmentation on the enhanced image data to obtain portrait area image data and background area image data;
constructing an image feature recognition model based on a single-layer convolutional neural network;
importing portrait region image data and background region image data into the image feature identification model for image feature extraction to obtain portrait feature data and background feature data;
performing image characteristic anomaly analysis on the portrait characteristic data and the background characteristic data to obtain portrait data anomaly information and background data anomaly information; information combination is carried out on the portrait data abnormal information and the background data abnormal information to obtain data abnormal classification information;
the image processing and data updating recovery based on the convolutional neural network are carried out according to the data anomaly classification information and the transmission image data to obtain updated transmission image data, and the method specifically comprises the following steps:
acquiring portrait data abnormal information in the data abnormal classification information;
acquiring a target object retrieval tag, and retrieving from a historical comparison image database according to the target object retrieval tag to obtain target object comparison image data;
performing portrait characteristic analysis on the contrast image data to obtain contrast portrait characteristics;
according to the portrait data abnormal information and the contrast portrait characteristics, performing data abnormal recovery processing on the portrait characteristic data to obtain second portrait characteristic data;
wherein, the image processing and data updating recovery based on the convolution neural network are carried out according to the data abnormal classification information and the transmission image data to obtain the updated transmission image data, and the method also comprises the following steps:
acquiring background data abnormal information in the data abnormal classification information;
acquiring background characteristic data, and performing characteristic retrieval on the background characteristic data from image big data to obtain retrieval characteristic data higher than preset similarity;
according to the background data abnormal information, in combination with the retrieval characteristic data, performing data abnormal recovery processing on the background characteristic data to obtain second background characteristic data;
and according to the second portrait characteristic data and the second background characteristic data, carrying out image abnormity correction on the transmission image data to obtain updated transmission image data.
2. The method for acquiring the on-line learning image based on the deep learning as claimed in claim 1, wherein the sampling of the on-line learning image according to the preset sampling rate to obtain the initial image data of the target object comprises:
according to a preset sampling rate, carrying out image sampling from a user terminal to obtain target object reference image data;
performing data anomaly analysis and screening on the target object reference image data to obtain non-anomalous image data;
carrying out image feature analysis on the non-abnormal image data, screening out image data with the image feature degree larger than a preset feature degree, and obtaining high-feature-contrast image data;
and constructing a history contrast image database, and importing the high-feature contrast image data into the history contrast image database.
3. The method for acquiring the on-line learning image based on the deep learning as claimed in claim 1, wherein the initial image data is subjected to data preprocessing according to a preset transmission standard to obtain transmission image data, and specifically comprises:
acquiring initial image data, and judging the storage format of the initial image data;
if the initial image data storage format is not the preset image storage format, performing lossless conversion on the preset image storage format to obtain converted image data;
and performing image data geometric transformation processing on the converted image data according to the preset image size and resolution to obtain transmission image data.
4. The method for collecting the on-line learning image based on the deep learning as claimed in claim 1, wherein the encrypting and compressing the updated transmission image data to transmit to a preset server specifically comprises:
performing matrix conversion on the updated and transmitted image data to obtain an image matrix;
dividing data according to the size of the image matrix to obtain N sub-matrixes;
carrying out data encryption on the data corresponding to the N sub-matrixes according to a preset encryption algorithm to obtain N encrypted data blocks;
merging and compressing the N encrypted data blocks to obtain result transmission data;
and sending the result transmission data to a preset server for data storage.
5. An on-line learning image acquisition system based on deep learning, characterized in that the system comprises: the on-line learning image acquisition method based on the deep learning is implemented by the processor, and the following steps are implemented when the on-line learning image acquisition method based on the deep learning is executed by the processor:
performing online learning image sampling according to a preset sampling rate to obtain initial image data of a target object;
carrying out data preprocessing on the initial image data according to a preset transmission standard to obtain transmission image data;
carrying out data anomaly analysis and classification on the transmitted image data to obtain data anomaly classification information;
performing image processing and data updating recovery based on a convolutional neural network according to the data anomaly classification information and the transmission image data to obtain updated transmission image data;
encrypting, compressing and transmitting the updated and transmitted image data to a preset server;
the method comprises the following steps of performing data anomaly analysis and classification on transmission image data to obtain data anomaly classification information, and specifically comprises the following steps:
carrying out image smoothing and noise reduction preprocessing on the transmission image data to obtain enhanced image data;
carrying out portrait area identification and image segmentation on the enhanced image data to obtain portrait area image data and background area image data;
constructing an image feature recognition model based on a single-layer convolution neural network;
importing the image data of the portrait area and the image data of the background area into the image feature recognition model for image feature extraction to obtain portrait feature data and background feature data;
performing image characteristic anomaly analysis on the portrait characteristic data and the background characteristic data to obtain portrait data anomaly information and background data anomaly information; information combination is carried out on the portrait data abnormal information and the background data abnormal information to obtain data abnormal classification information;
the image processing and data updating recovery based on the convolutional neural network are carried out according to the data anomaly classification information and the transmission image data to obtain updated transmission image data, and the method specifically comprises the following steps:
acquiring portrait data abnormal information in the data abnormal classification information;
acquiring a target object retrieval tag, and retrieving from a history comparison image database according to the target object retrieval tag to obtain target object comparison image data;
carrying out portrait feature analysis on the contrast image data to obtain contrast portrait features;
according to the portrait data abnormal information and the contrast portrait characteristics, performing data abnormal recovery processing on the portrait characteristic data to obtain second portrait characteristic data;
wherein, the image processing and data updating recovery based on the convolution neural network are carried out according to the data abnormal classification information and the transmission image data to obtain the updated transmission image data, and the method also comprises the following steps:
acquiring background data abnormal information in the data abnormal classification information;
acquiring background characteristic data, and performing characteristic retrieval on the background characteristic data from image big data to obtain retrieval characteristic data higher than preset similarity;
according to the background data abnormal information, in combination with the retrieval characteristic data, performing data abnormal recovery processing on the background characteristic data to obtain second background characteristic data;
and according to the second portrait characteristic data and the second background characteristic data, carrying out image abnormity correction on the transmission image data to obtain updated transmission image data.
6. The system of claim 5, wherein the sampling of the on-line learning image according to the preset sampling rate to obtain the initial image data of the target object comprises:
according to a preset sampling rate, carrying out image sampling from a user terminal to obtain target object reference image data;
performing data anomaly analysis and screening on the target object reference image data to obtain non-anomalous image data;
carrying out image characteristic analysis on the non-abnormal image data, screening out image data with the image characteristic degree larger than a preset characteristic degree, and obtaining high-characteristic contrast image data;
and constructing a history contrast image database, and importing the high-feature contrast image data into the history contrast image database.
7. The system for acquiring on-line learning images based on deep learning according to claim 5, wherein the image data to be updated and transmitted is encrypted, compressed and transmitted to a preset server, specifically:
performing matrix conversion on the updated and transmitted image data to obtain an image matrix;
dividing data according to the size of the image matrix to obtain N sub-matrixes;
carrying out data encryption on the data corresponding to the N sub-matrixes according to a preset encryption algorithm to obtain N encrypted data blocks;
merging and compressing the N encrypted data blocks to obtain result transmission data;
and sending the result transmission data to a preset server for data storage.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211128260.8A CN115205952B (en) | 2022-09-16 | 2022-09-16 | Online learning image acquisition method and system based on deep learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211128260.8A CN115205952B (en) | 2022-09-16 | 2022-09-16 | Online learning image acquisition method and system based on deep learning |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115205952A CN115205952A (en) | 2022-10-18 |
| CN115205952B true CN115205952B (en) | 2022-11-25 |
Family
ID=83573195
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211128260.8A Active CN115205952B (en) | 2022-09-16 | 2022-09-16 | Online learning image acquisition method and system based on deep learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115205952B (en) |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108259848A (en) * | 2018-04-04 | 2018-07-06 | 武汉市技领科技有限公司 | A kind of identification device and monitoring device |
| CN111565303A (en) * | 2020-05-29 | 2020-08-21 | 深圳市易链信息技术有限公司 | Video monitoring method, system and readable storage medium based on fog calculation and deep learning |
| CN112711766A (en) * | 2021-01-05 | 2021-04-27 | 重庆第二师范学院 | Image transmission system based on scrambling block compressed sensing, encryption method and decryption method |
| CN112838922A (en) * | 2021-01-22 | 2021-05-25 | 广东工业大学 | DICOM image asymmetric encryption method based on chaotic mapping and selective Signcryption |
| CN112887587A (en) * | 2021-01-08 | 2021-06-01 | 浙江工贸职业技术学院 | Self-adaptive image data fast transmission method capable of carrying out wireless connection |
| CN114143551A (en) * | 2021-12-01 | 2022-03-04 | 北京邮电大学 | Video safe and efficient transmission system applied to video sensor network |
| CN114979407A (en) * | 2022-05-24 | 2022-08-30 | 浙江科技学院 | Multi-image encryption and decryption method based on code division multiple access and deep learning ghost imaging |
Family Cites Families (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101090497A (en) * | 2007-06-26 | 2007-12-19 | 北京中星微电子有限公司 | Data compression method and device |
| CN103366386A (en) * | 2013-07-14 | 2013-10-23 | 西安电子科技大学 | Parallel image uncompressing system based on multiple processes and multiple threads |
| CN106254242A (en) * | 2016-08-04 | 2016-12-21 | 胡汉强 | A kind of data transmission method, Centralized Controller, forwarding face equipment and local terminal communicator |
| CN106713816A (en) * | 2017-01-09 | 2017-05-24 | 宇龙计算机通信科技(深圳)有限公司 | Image transmission method and device |
| CN107066411A (en) * | 2017-04-13 | 2017-08-18 | 深圳市酷开网络科技有限公司 | Data transmission method, device and computer-readable recording medium |
| CN107948472A (en) * | 2017-11-29 | 2018-04-20 | 北京星云环影科技有限责任公司 | Intelligent monitoring pick-up head and safety-protection system |
| CN108449571A (en) * | 2018-01-30 | 2018-08-24 | 深圳市元征软件开发有限公司 | A kind of car monitoring method and equipment |
| JP6337226B1 (en) * | 2018-03-02 | 2018-06-06 | 株式会社エネルギア・コミュニケーションズ | Abnormal point detection system |
| CN110232719B (en) * | 2019-06-21 | 2021-10-26 | 腾讯科技（深圳）有限公司 | Medical image classification method, model training method and server |
| CN111131655A (en) * | 2019-12-12 | 2020-05-08 | 成都信息工程大学 | High-speed fault-tolerant image transmission method based on pest and disease damage research |
| CN111062323B (en) * | 2019-12-16 | 2023-06-02 | 腾讯科技（深圳）有限公司 | Face image transmission method, numerical value transfer method, device and electronic equipment |
| WO2021119982A1 (en) * | 2019-12-17 | 2021-06-24 | 深圳市大疆创新科技有限公司 | Image transmission system and method, control apparatus, and mobile platform |
| CN114793146A (en) * | 2021-01-26 | 2022-07-26 | 普联技术有限公司 | Data receiving method, data sending method and related equipment |
| CN113746617B (en) * | 2021-08-18 | 2023-07-18 | 暨南大学 | Image transmission and reconstruction method and device based on chaotic compression encryption |
| CN113923315B (en) * | 2021-12-13 | 2022-03-29 | 湖南菠萝互娱网络信息有限公司 | Image cloud storage leakage-prevention encryption system |
| CN114821240A (en) * | 2022-05-11 | 2022-07-29 | 网易（杭州）网络有限公司 | Abnormal image detection method and device, electronic equipment and storage medium |
-
2022
- 2022-09-16 CN CN202211128260.8A patent/CN115205952B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108259848A (en) * | 2018-04-04 | 2018-07-06 | 武汉市技领科技有限公司 | A kind of identification device and monitoring device |
| CN111565303A (en) * | 2020-05-29 | 2020-08-21 | 深圳市易链信息技术有限公司 | Video monitoring method, system and readable storage medium based on fog calculation and deep learning |
| CN112711766A (en) * | 2021-01-05 | 2021-04-27 | 重庆第二师范学院 | Image transmission system based on scrambling block compressed sensing, encryption method and decryption method |
| CN112887587A (en) * | 2021-01-08 | 2021-06-01 | 浙江工贸职业技术学院 | Self-adaptive image data fast transmission method capable of carrying out wireless connection |
| CN112838922A (en) * | 2021-01-22 | 2021-05-25 | 广东工业大学 | DICOM image asymmetric encryption method based on chaotic mapping and selective Signcryption |
| CN114143551A (en) * | 2021-12-01 | 2022-03-04 | 北京邮电大学 | Video safe and efficient transmission system applied to video sensor network |
| CN114979407A (en) * | 2022-05-24 | 2022-08-30 | 浙江科技学院 | Multi-image encryption and decryption method based on code division multiple access and deep learning ghost imaging |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115205952A (en) | 2022-10-18 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Manap et al. | Non-distortion-specific no-reference image quality assessment: A survey | |
| CN108305238B (en) | Image processing method, image processing device, storage medium and computer equipment | |
| Xue et al. | Reduced reference image quality assessment based on Weibull statistics | |
| Corchs et al. | No reference image quality classification for JPEG-distorted images | |
| CN109872305B (en) | No-reference stereo image quality evaluation method based on quality map generation network | |
| US20210192019A1 (en) | System and method for digital steganography purification | |
| CN115302963A (en) | Bar code printing control method, system and medium based on machine vision | |
| CN111179202A (en) | Single image defogging enhancement method and system based on generation countermeasure network | |
| Shaik et al. | A review of hashing based image authentication techniques | |
| CN113705426A (en) | Face checking method, device, server and readable storage medium | |
| CN112395503A (en) | Face recognition-based sharing platform intelligent recommendation method and system and readable storage medium | |
| CN115861210A (en) | Transformer substation equipment abnormity detection method and system based on twin network | |
| CN113792659B (en) | Document identification method and device and electronic equipment | |
| CN109785286B (en) | Image restoration detection method based on texture feature fusion | |
| CN115205952B (en) | Online learning image acquisition method and system based on deep learning | |
| US6487311B1 (en) | OCR-based image compression | |
| CN115695988A (en) | Intelligent equipment true random key generation method, device, terminal and medium | |
| CN110213457B (en) | Image transmission method and device | |
| CN109101646B (en) | Data processing method, device, system and computer readable medium | |
| CN113723466B (en) | Text similarity quantification method, device and system | |
| CN116471210B (en) | Node penetration monitoring method, system, equipment and readable storage medium | |
| CN112669240B (en) | High-definition image restoration method and device, electronic equipment and storage medium | |
| Charrier et al. | Dimension reduction-based attributes selection in no-reference learning-based image quality algorithms | |
| CN112200247B (en) | Image processing system and method based on multi-dimensional image mapping | |
| CN113447572B (en) | Steel rail flaw detection method, electronic device, steel rail flaw detection vehicle and readable storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |