CN109089010A - A kind of image transfer method and device - Google Patents
A kind of image transfer method and device Download PDFInfo
- Publication number
- CN109089010A CN109089010A CN201811074339.0A CN201811074339A CN109089010A CN 109089010 A CN109089010 A CN 109089010A CN 201811074339 A CN201811074339 A CN 201811074339A CN 109089010 A CN109089010 A CN 109089010A
- Authority
- CN
- China
- Prior art keywords
- image
- feature value
- value parameter
- neural network
- described image
- 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.)
- Pending
Links
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N1/00—Scanning, transmission or reproduction of documents or the like, e.g. facsimile transmission; Details thereof
- H04N1/00095—Systems or arrangements for the transmission of the picture signal
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N1/00—Scanning, transmission or reproduction of documents or the like, e.g. facsimile transmission; Details thereof
- H04N1/40—Picture signal circuits
Abstract
The present invention relates to technical field of image processing, a kind of image transfer method and device are disclosed, this method comprises: obtaining image to be transmitted, and described image is sent to processor；Positive processing is carried out to described image, extracts feature value parameter；The feature value parameter is transmitted to server；Reverse process is carried out to the feature value parameter, reduction obtains described image；Characteristic value by extracting image to be transmitted carries out image transmitting, reduces transmission data, improves the speed of transmission, reduce network delay.
Description
Technical field
The present invention relates to technical field of image processing more particularly to a kind of image transfer methods and device.
Background technique
With the development of network technology in recent years, image transfer through networks system starts to be widely applied, and this system passes through net
Network transmits image, in general, if image data amount is very greatly.Can exist and be passed caused by network delay and network load
Defeated speed is slow, and picture quality cannot ensure.Usually first then compression image information passes through communication network to conventional images transmission technology
Network is transferred to terminal, the unstable generation network delay of meeting.A large amount of image data causes network load to pass by network transmission
It is defeated very slow, and then the problems such as will lead to image fault.
Summary of the invention
It is a primary object of the present invention to propose a kind of image transfer method and device, pass through the spy for extracting image to be transmitted
Value indicative carries out image transmitting, reduces transmission data, improves the speed of transmission, reduce network delay.
To achieve the above object, a kind of image transfer method provided by the invention, comprising:
Image to be transmitted is obtained, and described image is sent to processor；
Positive processing is carried out to described image, extracts feature value parameter；
The feature value parameter is transmitted to server；
Reverse process is carried out to the feature value parameter, reduction obtains described image.
Optionally, described that positive processing is carried out to described image, before extracting feature value parameter further include:
Neural network is trained, trained neural network is obtained, obtains the corresponding weight of each layer of neural network
And Reduction parameter.
Optionally, described to carry out positive processing to described image, extracting feature value parameter includes:
Positive processing is carried out to described image by neural network deep learning model；
The feature value parameter of described image is extracted by the corresponding weight of each layer of neural network.
Optionally, described to carry out reverse process to the feature value parameter, reduction obtains described image specifically:
Reverse process is carried out to the feature value parameter by each layer of neural network corresponding Reduction parameter, reduction obtains
Described image.
It is optionally, described that the feature value parameter is transmitted to server specifically:
The feature value parameter is transmitted to server by wireless network.
As another aspect of the present invention, a kind of image transmission for providing, comprising:
Image collection module is sent to processor for obtaining image to be transmitted, and by described image；
Positive processing module extracts feature value parameter for carrying out positive processing to described image；
Transmission module, for the feature value parameter to be transmitted to server；
Recovery module, for carrying out reverse process to the feature value parameter, reduction obtains described image.
Optionally, further includes:
Training module obtains trained neural network, obtains each layer of nerve net for being trained to neural network
The corresponding weight of network and Reduction parameter.
Optionally, the positive processing module includes:
Unit, for carrying out positive processing to described image by neural network deep learning model；
Extraction unit, the characteristic value for extracting described image by the corresponding weight of each layer of neural network are joined
Number.
Optionally, the recovery module specifically:
Reverse process is carried out to the feature value parameter by each layer of neural network corresponding Reduction parameter, reduction obtains
Described image.
Optionally, the transmission module specifically:
The feature value parameter is transmitted to server by wireless network.
A kind of image transfer method and device proposed by the present invention, this method comprises: obtain image to be transmitted, and by institute
It states image and is sent to processor；Positive processing is carried out to described image, extracts feature value parameter；The feature value parameter is transmitted
To server；Reverse process is carried out to the feature value parameter, reduction obtains described image；By the spy for extracting image to be transmitted
Value indicative carries out image transmitting, reduces transmission data, improves the speed of transmission, reduce network delay.
Detailed description of the invention
Fig. 1 is a kind of flow chart for image transfer method that the embodiment of the present invention one provides；
Fig. 2 is the flow chart for another image transfer method that the embodiment of the present invention one provides；
Fig. 3 is the method flow diagram of step S20 in Fig. 1；
Fig. 4 is the positive processing and inverse process schematic diagram that the embodiment of the present invention one provides；
Fig. 5 is a kind of exemplary block diagram of image transmission provided by Embodiment 2 of the present invention；
Fig. 6 is the exemplary block diagram of another image transmission provided by Embodiment 2 of the present invention；
Fig. 7 is the exemplary block diagram of positive processing module in Fig. 5.
The embodiments will be further described with reference to the accompanying drawings for the realization, the function and the advantages of the object of the present invention.
Specific embodiment
It should be appreciated that the specific embodiments described herein are merely illustrative of the present invention, it is not intended to limit the present invention.
In subsequent description, it is only using the suffix for indicating such as " module ", " component " or " unit " of element
Be conducive to explanation of the invention, itself there is no specific meanings.Therefore, " module " can mixedly make with " component "
With.
Embodiment
As shown in Figure 1, in the present embodiment, a kind of image transfer method, comprising:
S10, image to be transmitted is obtained, and described image is sent to processor；
S20, positive processing is carried out to described image, extracts feature value parameter；
S30, the feature value parameter is transmitted to server；
S40, reverse process is carried out to the feature value parameter, reduction obtains described image.
In the present embodiment, image transmitting is carried out by extracting the characteristic value of image to be transmitted, reduces transmission data, mentions
The high speed of transmission, reduces network delay.
In the present embodiment, picture or image information are acquired by photographic device, is sent to place as image to be transmitted
Device is managed, processor receives the image of photographic device upload, carries out just by neural network deep learning model to described image
To processing, feature value parameter is extracted, compared with the initial data of image, characteristic value supplemental characteristic amount is smaller, therefore, improves
The speed of data transmission.
As shown in Fig. 2, in the present embodiment, before the step S20 further include:
S11, neural network is trained, obtains trained neural network, it is corresponding to obtain each layer of neural network
Weight and Reduction parameter.
In the present embodiment, by the way that a certain number of image patterns are inputted neural network one by one, it is trained so that about
The preset function value of all image patterns is minimum, i.e., neural network has reached convergence, has thus just obtained trained nerve net
Network.In the training process, X(k+1)=g (WkXk), wherein WkFor the corresponding weight of each layer of neural network, g is receptance function,
X(k)=(Xk1, Xk2..., Xkn), and be ∑ (X about the preset function of all image patterns(n)-y)2, y is the matter of image pattern
Magnitude, that is, the preset function be it is squared to the input and output difference of all image patterns and, result in every layer of WkIt is corresponding
Weight (W1,W2..., Wn) and every layer of AxReduction parameter (A1,A2..., AX),To obtain trained nerve net
Network.
As shown in figure 3, in the present embodiment, the step S20 includes:
S21, positive processing is carried out to described image by neural network deep learning model；
S22, the feature value parameter that described image is extracted by the corresponding weight of each layer of neural network.
In the present embodiment, the step S40 specifically:
Reverse process is carried out to the feature value parameter by each layer of neural network corresponding Reduction parameter, reduction obtains
Described image.
In the present embodiment, reverse process, every layer of corresponding equity value W are carried out by deep learning modelK(W1,
W2..., Wn) pass through every layer of AnReduction parameter (A1,A2..., An)Processing, obtains original image X.
It is in the present embodiment, described that the feature value parameter is transmitted to server specifically:
The feature value parameter is transmitted to server by wireless network.
As shown in figure 4, including during reverse process for the positive processing of the present invention and inverse process schematic diagram
Using obtained judging result as a parameter A in backpropagation approach, the preset function ∑ (X of image pattern is used(n)-y)2
The error for finding out a value code calculated value and true value, the value for being calculated the cost function according to chain rule is to upper one layer
Each weight carry out chain type derivation, find out come derivative as new weight, cost function is calculated after having sought this layer of weight
Value out to upper one layer of progress derivation again, and so on, until each layer all completes derivation and obtains corresponding weighted value.
Embodiment two
As shown in figure 5, in the present embodiment, a kind of image transmission, comprising:
Image collection module 10 is sent to processor for obtaining image to be transmitted, and by described image；
Positive processing module 20 extracts feature value parameter for carrying out positive processing to described image；
Transmission module 30, for the feature value parameter to be transmitted to server；
Recovery module 40, for carrying out reverse process to the feature value parameter, reduction obtains described image.
In the present embodiment, image transmitting is carried out by extracting the characteristic value of image to be transmitted, reduces transmission data, mentions
The high speed of transmission, reduces network delay.
In the present embodiment, picture or image information are acquired by photographic device, is sent to place as image to be transmitted
Device is managed, processor receives the image of photographic device upload, carries out just by neural network deep learning model to described image
To processing, feature value parameter is extracted, compared with the initial data of image, characteristic value supplemental characteristic amount is smaller, therefore, improves
The speed of data transmission.
As shown in fig. 6, in the present embodiment, a kind of image transmission further include:
Training module 50 obtains trained neural network, obtains each layer of nerve for being trained to neural network
The corresponding weight of network and Reduction parameter.
In the present embodiment, by the way that a certain number of image patterns are inputted neural network one by one, it is trained so that about
The preset function value of all image patterns is minimum, i.e., neural network has reached convergence, has thus just obtained trained nerve net
Network.In the training process, X(k+1)=g (WkXk), wherein WkFor the corresponding weight of each layer of neural network, g is receptance function,
X(k)=(Xk1, Xk2..., Xkn), and be ∑ (X about the preset function of all image patterns(n)-y)2, y is the matter of image pattern
Magnitude, that is, the preset function be it is squared to the input and output difference of all image patterns and, result in every layer of WkIt is corresponding
Weight (W1, W2..., Wn) and every layer of AxReduction parameter (A1,A2..., AX),To obtain trained nerve net
Network.
As shown in fig. 7, in the present embodiment, the forward direction processing module includes:
Unit 21, for carrying out positive processing to described image by neural network deep learning model；
Extraction unit 22, the characteristic value for extracting described image by the corresponding weight of each layer of neural network are joined
Number.
In the present embodiment, the recovery module specifically:
Reverse process is carried out to the feature value parameter by each layer of neural network corresponding Reduction parameter, reduction obtains
Described image.
In the present embodiment, reverse process, every layer of corresponding equity value W are carried out by deep learning modelK(W1,
W2..., Wn) pass through every layer of AnReduction parameter (A1,A2..., An)Processing, obtains original image X.
It is in the present embodiment, described that the feature value parameter is transmitted to server specifically:
The feature value parameter is transmitted to server by wireless network.
As shown in figure 4, including during reverse process for the positive processing of the present invention and inverse process schematic diagram
Using obtained judging result as a parameter A in backpropagation approach, the preset function ∑ (X of image pattern is used(n)-y)2
The error for finding out a value code calculated value and true value, the value for being calculated the cost function according to chain rule is to upper one layer
Each weight carry out chain type derivation, find out come derivative as new weight, cost function is calculated after having sought this layer of weight
Value out to upper one layer of progress derivation again, and so on, until each layer all completes derivation and obtains corresponding weighted value.
In the present embodiment, the transmission module specifically:
The feature value parameter is transmitted to server by wireless network.
It should be noted that, in this document, the terms "include", "comprise" or its any other variant are intended to non-row
His property includes, so that the process, method, article or the device that include a series of elements not only include those elements, and
And further include other elements that are not explicitly listed, or further include for this process, method, article or device institute it is intrinsic
Element.In the absence of more restrictions, the element limited by sentence "including a ...", it is not excluded that including being somebody's turn to do
There is also other identical elements in the process, method of element, article or device.
The serial number of the above embodiments of the invention is only for description, does not represent the advantages or disadvantages of the embodiments.
Through the above description of the embodiments, those skilled in the art can be understood that above-described embodiment side
Method can be realized by means of software and necessary general hardware platform, naturally it is also possible to by hardware, but in many cases
The former is more preferably embodiment.Based on this understanding, technical solution of the present invention substantially in other words does the prior art
The part contributed out can be embodied in the form of software products, which is stored in a storage medium
In (such as ROM/RAM, magnetic disk, CD), including some instructions are used so that a terminal device (can be mobile phone, computer, clothes
Business device, air conditioner or the network equipment etc.) execute method described in each embodiment of the present invention.
The above is only a preferred embodiment of the present invention, is not intended to limit the scope of the invention, all to utilize this hair
Equivalent structure or equivalent flow shift made by bright specification and accompanying drawing content is applied directly or indirectly in other relevant skills
Art field, is included within the scope of the present invention.
Claims (10)
1. a kind of image transfer method characterized by comprising
Image to be transmitted is obtained, and described image is sent to processor；
Positive processing is carried out to described image, extracts feature value parameter；
The feature value parameter is transmitted to server；
Reverse process is carried out to the feature value parameter, reduction obtains described image.
2. a kind of image transfer method according to claim 1, which is characterized in that described to be carried out at forward direction to described image
It manages, before extraction feature value parameter further include:
Neural network is trained, trained neural network is obtained, obtain the corresponding weight of each layer of neural network and is gone back
Original parameter.
3. a kind of image transfer method according to claim 2, which is characterized in that described to be carried out at forward direction to described image
Reason, extracting feature value parameter includes:
Positive processing is carried out to described image by neural network deep learning model；
The feature value parameter of described image is extracted by the corresponding weight of each layer of neural network.
4. a kind of image transfer method according to claim 2, which is characterized in that described to be carried out to the feature value parameter
Reverse process, reduction obtain described image specifically:
Reverse process is carried out to the feature value parameter by each layer of neural network corresponding Reduction parameter, reduction obtains described
Image.
5. a kind of image transfer method according to claim 1, which is characterized in that described to transmit the feature value parameter
To server specifically:
The feature value parameter is transmitted to server by wireless network.
6. a kind of image transmission characterized by comprising
Image collection module is sent to processor for obtaining image to be transmitted, and by described image；
Positive processing module extracts feature value parameter for carrying out positive processing to described image；
Transmission module, for the feature value parameter to be transmitted to server；
Recovery module, for carrying out reverse process to the feature value parameter, reduction obtains described image.
7. a kind of image transmission according to claim 6, which is characterized in that further include:
Training module obtains trained neural network, obtains each layer of neural network pair for being trained to neural network
The weight and Reduction parameter answered.
8. a kind of image transmission according to claim 7, which is characterized in that it is described forward direction processing module include:
Unit, for carrying out positive processing to described image by neural network deep learning model；
Extraction unit, for extracting the feature value parameter of described image by the corresponding weight of each layer of neural network.
9. a kind of image transmission according to claim 7, which is characterized in that the recovery module specifically:
Reverse process is carried out to the feature value parameter by each layer of neural network corresponding Reduction parameter, reduction obtains described
Image.
10. a kind of image transmission according to claim 6, which is characterized in that the transmission module specifically:
The feature value parameter is transmitted to server by wireless network.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811074339.0A CN109089010A (en) | 2018-09-14 | 2018-09-14 | A kind of image transfer method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811074339.0A CN109089010A (en) | 2018-09-14 | 2018-09-14 | A kind of image transfer method and device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109089010A true CN109089010A (en) | 2018-12-25 |
Family
ID=64841432
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811074339.0A Pending CN109089010A (en) | 2018-09-14 | 2018-09-14 | A kind of image transfer method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109089010A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110213061A (en) * | 2019-05-23 | 2019-09-06 | 北京力拓飞远科技有限公司 | Synchronized communication method, synchronized communication means, synchronous communication equipment and medium |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105550701A (en) * | 2015-12-09 | 2016-05-04 | 福州华鹰重工机械有限公司 | Real-time image extraction and recognition method and device |
| US20170220923A1 (en) * | 2016-02-02 | 2017-08-03 | Samsung Electronics Co., Ltd. | Gesture classification apparatus and method using emg signal |
| CN107256414A (en) * | 2017-06-27 | 2017-10-17 | 哈尔滨工业大学 | Polarimetric SAR Image convolutional neural networks sorting technique based on spatial feature collection of illustrative plates |
| CN108062780A (en) * | 2017-12-29 | 2018-05-22 | 百度在线网络技术（北京）有限公司 | Method for compressing image and device |
-
2018
- 2018-09-14 CN CN201811074339.0A patent/CN109089010A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105550701A (en) * | 2015-12-09 | 2016-05-04 | 福州华鹰重工机械有限公司 | Real-time image extraction and recognition method and device |
| US20170220923A1 (en) * | 2016-02-02 | 2017-08-03 | Samsung Electronics Co., Ltd. | Gesture classification apparatus and method using emg signal |
| CN107256414A (en) * | 2017-06-27 | 2017-10-17 | 哈尔滨工业大学 | Polarimetric SAR Image convolutional neural networks sorting technique based on spatial feature collection of illustrative plates |
| CN108062780A (en) * | 2017-12-29 | 2018-05-22 | 百度在线网络技术（北京）有限公司 | Method for compressing image and device |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110213061A (en) * | 2019-05-23 | 2019-09-06 | 北京力拓飞远科技有限公司 | Synchronized communication method, synchronized communication means, synchronous communication equipment and medium |
| CN110213061B (en) * | 2019-05-23 | 2020-06-05 | 北京力拓飞远科技有限公司 | Synchronous communication method, synchronous communication device, synchronous communication apparatus, and medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111310932A (en) | Method, device and equipment for optimizing horizontal federated learning system and readable storage medium | |
| WO2020098422A1 (en) | Encoded pattern processing method and device , storage medium and electronic device | |
| CN105095919A (en) | Image recognition method and image recognition device | |
| US20220067888A1 (en) | Image processing method and apparatus, storage medium, and electronic device | |
| CN103299342B (en) | Method and apparatus for providing a mechanism for gesture recognition | |
| CN112001274B (en) | Crowd density determining method, device, storage medium and processor | |
| CN110598982B (en) | Active wind control method and system based on intelligent interaction | |
| CN112906721B (en) | Image processing method, device, equipment and computer readable storage medium | |
| JP2023518015A (en) | Data protection method, device, server and medium | |
| Chen et al. | Naturalization module in neural networks for screen content image quality assessment | |
| WO2020062191A1 (en) | Image processing method, apparatus and device | |
| CN111047543A (en) | Image enhancement method, device and storage medium | |
| CN111985281A (en) | Image generation model generation method and device and image generation method and device | |
| CN115424088A (en) | Image processing model training method and device | |
| CN115064020A (en) | Intelligent teaching method, system and storage medium based on digital twin technology | |
| CN110795558B (en) | Label acquisition method and device, storage medium and electronic device | |
| CN112084959A (en) | Crowd image processing method and device | |
| KR20210116922A (en) | Method and Device for Fast Adaptation through Meta-learning of Super Resolution Model | |
| CN109089010A (en) | A kind of image transfer method and device | |
| CN110489955B (en) | Image processing, device, computing device and medium applied to electronic equipment | |
| CN110490876B (en) | Image segmentation method based on lightweight neural network | |
| CN112115703B (en) | Article evaluation method and device | |
| CN111553961B (en) | Method and device for acquiring line manuscript corresponding color map, storage medium and electronic device | |
| JP2021120840A (en) | Learning method, device, and program | |
| CN111402133A (en) | Image processing method, image processing device, electronic equipment and computer readable 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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20181225 |