CN115206230A - Drive circuit and drive control method thereof - Google Patents
Drive circuit and drive control method thereof Download PDFInfo
- Publication number
- CN115206230A CN115206230A CN202210909527.0A CN202210909527A CN115206230A CN 115206230 A CN115206230 A CN 115206230A CN 202210909527 A CN202210909527 A CN 202210909527A CN 115206230 A CN115206230 A CN 115206230A
- Authority
- CN
- China
- Prior art keywords
- matrix
- characteristic
- feature
- vector
- real
- 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.)
- Granted
Links
Images
Classifications
-
- G—PHYSICS
- G09—EDUCATION; CRYPTOGRAPHY; DISPLAY; ADVERTISING; SEALS
- G09G—ARRANGEMENTS OR CIRCUITS FOR CONTROL OF INDICATING DEVICES USING STATIC MEANS TO PRESENT VARIABLE INFORMATION
- G09G3/00—Control arrangements or circuits, of interest only in connection with visual indicators other than cathode-ray tubes
- G09G3/20—Control arrangements or circuits, of interest only in connection with visual indicators other than cathode-ray tubes for presentation of an assembly of a number of characters, e.g. a page, by composing the assembly by combination of individual elements arranged in a matrix no fixed position being assigned to or needed to be assigned to the individual characters or partial characters
- G09G3/22—Control arrangements or circuits, of interest only in connection with visual indicators other than cathode-ray tubes for presentation of an assembly of a number of characters, e.g. a page, by composing the assembly by combination of individual elements arranged in a matrix no fixed position being assigned to or needed to be assigned to the individual characters or partial characters using controlled light sources
- G09G3/30—Control arrangements or circuits, of interest only in connection with visual indicators other than cathode-ray tubes for presentation of an assembly of a number of characters, e.g. a page, by composing the assembly by combination of individual elements arranged in a matrix no fixed position being assigned to or needed to be assigned to the individual characters or partial characters using controlled light sources using electroluminescent panels
- G09G3/32—Control arrangements or circuits, of interest only in connection with visual indicators other than cathode-ray tubes for presentation of an assembly of a number of characters, e.g. a page, by composing the assembly by combination of individual elements arranged in a matrix no fixed position being assigned to or needed to be assigned to the individual characters or partial characters using controlled light sources using electroluminescent panels semiconductive, e.g. using light-emitting diodes [LED]
-
- G—PHYSICS
- G09—EDUCATION; CRYPTOGRAPHY; DISPLAY; ADVERTISING; SEALS
- G09G—ARRANGEMENTS OR CIRCUITS FOR CONTROL OF INDICATING DEVICES USING STATIC MEANS TO PRESENT VARIABLE INFORMATION
- G09G2320/00—Control of display operating conditions
- G09G2320/06—Adjustment of display parameters
- G09G2320/0626—Adjustment of display parameters for control of overall brightness
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02B—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO BUILDINGS, e.g. HOUSING, HOUSE APPLIANCES OR RELATED END-USER APPLICATIONS
- Y02B20/00—Energy efficient lighting technologies, e.g. halogen lamps or gas discharge lamps
- Y02B20/40—Control techniques providing energy savings, e.g. smart controller or presence detection
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Computer Hardware Design (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Electroluminescent Light Sources (AREA)
- Control Of Indicators Other Than Cathode Ray Tubes (AREA)
Abstract
The application relates to the field of light emitting diode displays, and particularly discloses a driving circuit and a driving control method thereof, which are used for mining implicit associated characteristics of real-time driving voltage and real-time temperature values of all light emitting points respectively on the basis of a deep neural network model, and introducing topological characteristic information in order to pay attention to spatial association among the light emitting points, so that compensation voltage of each light emitting point can be determined on the basis of global characteristic information of the light emitting points, the problem of brightness attenuation of a display picture is solved through the compensation voltage, and the AMLED can display a high-quality picture. In this way, the panel light emission consistency and stability of the display can be optimized.
Description
Technical Field
The present invention relates to the field of light emitting diode displays, and more particularly, to a driving circuit and a driving control method thereof.
Background
An Active Micro-LED (AMLED) display has the advantages of high contrast, high color saturation, high brightness, etc., and thus becomes one of the next generation hot display technologies. When the conventional AMLED display is operated, the temperature of the panel increases, so that the thermal effect of the light emitting diode, that is, the light emitting efficiency and the voltage (Vled) thereof become small, and the brightness of the AMLED display is attenuated.
Accordingly, it is desirable to solve the thermal effect of the light emitting diode and provide an AMLED display with high display quality.
Disclosure of Invention
The present application is proposed to solve the above-mentioned technical problems. The embodiment of the application provides a driving circuit and a driving control method thereof, which respectively perform implicit association feature mining on real-time driving voltage and real-time temperature values of all light-emitting points based on a deep neural network model, introduce topological feature information in order to pay attention to spatial association among the light-emitting points, and determine compensation voltage of each light-emitting point based on global feature information of the light-emitting points, so that the problem of brightness attenuation of a display picture is improved through the compensation voltage, and an AMLED can display a high-quality picture. In this way, the panel light emission consistency and stability of the display can be optimized.
According to an aspect of the present application, there is provided a driving circuit including:
the luminous point scanning module is used for acquiring real-time driving voltages and real-time temperature values of all luminous points in the display;
the spatial topology construction module is used for constructing a topology matrix between a light-emitting point to be adjusted in the display and all other light-emitting points in the display, wherein the characteristic value of each position in the topology matrix is the distance between the light-emitting point to be adjusted and each other light-emitting point;
the structuralization module is used for constructing the real-time driving voltages of all the light-emitting points in the display into a driving voltage input matrix and constructing the real-time temperature values of all the light-emitting points in the display into a temperature input matrix;
the encoding module is used for enabling the driving voltage input matrix and the temperature input matrix to pass through a first convolution neural network serving as a feature extractor respectively so as to obtain a driving voltage feature matrix and a temperature feature matrix;
the first fusion module is used for fusing the driving voltage characteristic matrix and the temperature characteristic matrix to obtain a superposition characteristic matrix;
the spatial topological coding module is used for enabling the topological matrix to pass through a second convolutional neural network serving as a feature extractor so as to obtain a topological feature matrix;
the second fusion module is used for multiplying the superposition characteristic matrix and the topological characteristic matrix by a matrix, and mapping the high-dimensional topological information of the topological characteristic matrix to a high-dimensional characteristic space of the superposition characteristic matrix to obtain an influence characteristic matrix based on space topology;
the current state coding module is used for enabling the real-time driving voltage and the real-time temperature value of the luminous point to be adjusted to pass through a context coder comprising an embedded layer so as to obtain a current state vector;
the mapping module is used for multiplying the current state vector by the influence characteristic matrix to obtain a decoding characteristic vector;
the characteristic distribution correction module is used for carrying out characteristic distribution correction on the influence characteristic matrix to obtain an enhanced generalization expression vector;
the optimization module is used for multiplying the enhanced generalization expression vector and the decoding characteristic vector according to position points to obtain an enhanced decoding characteristic vector; and
and the compensation voltage value generation module is used for decoding and regressing the enhanced decoding characteristic vector through a decoder to obtain a decoding value, and the decoding value is a compensation voltage value.
In the above driving circuit, the encoding module is further configured to: each layer of the first convolutional neural network performs in the forward pass of the layer: performing convolution processing on input data to obtain a convolution characteristic diagram; performing mean pooling based on local channel dimensions on the convolution feature map to obtain a pooled feature map; and performing nonlinear activation on the pooled feature map to obtain an activated feature map; wherein the outputs of the last layer of the first convolutional neural network are the driving voltage characteristic matrix and the temperature characteristic matrix, and the inputs of the first layer of the first convolutional neural network are the driving voltage input matrix and the temperature input matrix.
In the above driving circuit, the first fusing module is further configured to: fusing the driving voltage characteristic matrix and the temperature characteristic matrix according to the following formula to obtain the superposed characteristic matrix;
wherein the formula is:
M s ＝M 1 ⊙M 2
wherein, M s For the superimposed feature matrix, M 1 For said driving voltage characteristic matrix, M 2 As the temperature characteristic matrix, ", indicates dot-by-dot.
In the above driving circuit, the spatial topology coding module is further configured to: each layer of the second convolutional neural network performs in the forward pass of the layer: carrying out convolution processing on input data to obtain a convolution characteristic diagram; performing mean pooling based on local channel dimensions on the convolution feature map to obtain a pooled feature map; and performing nonlinear activation on the pooled feature map to obtain an activated feature map; wherein the output of the last layer of the second convolutional neural network is the topological feature matrix, and the input of the first layer of the second convolutional neural network is the topological matrix.
In the above driving circuit, the current state encoding module is further configured to: respectively converting the real-time driving voltage and the real-time temperature value of the luminous point to be regulated into input vectors by using the embedded layer of the context encoder containing the embedded layer so as to obtain a sequence of the input vectors; and globally context-based semantic-coding the sequence of input vectors using the converter including the context encoder of the embedding layer to obtain the current state vector.
In the above driving circuit, the characteristic distribution correction module is further configured to: performing characteristic distribution correction on the influence characteristic matrix by using the following formula to obtain the enhanced generalization expression vector;
wherein the formula is:
wherein m is i,j Is a mapping of the impact feature matrix to [0,1]Eigenvalues within the interval, and i m i,j representing the summation of said matrix of impact signatures along columns.
In the above driving circuit, the compensation voltage value generating module is further configured to: make itDecoding the enhanced decoded feature vector with a plurality of fully-connected layers of the decoder to obtain the decoded value by performing a decoding regression on the enhanced decoded feature vector with the following formula:
wherein X is the enhanced decoded feature vector, Y is the decoded value, W is a weight matrix, B is a bias vector,
representing the matrix multiplication, h (-) is the activation function.
According to another aspect of the present application, a driving control method of a driving circuit includes:
acquiring real-time driving voltages and real-time temperature values of all light-emitting points in a display;
constructing a topology matrix between a luminous point to be adjusted in the display and all other luminous points in the display, wherein the characteristic value of each position in the topology matrix is the distance between the luminous point to be adjusted and each other luminous point;
constructing real-time driving voltages of all light-emitting points in the display into a driving voltage input matrix and constructing real-time temperature values of all light-emitting points in the display into a temperature input matrix;
respectively passing the driving voltage input matrix and the temperature input matrix through a first convolution neural network serving as a feature extractor to obtain a driving voltage feature matrix and a temperature feature matrix;
fusing the driving voltage characteristic matrix and the temperature characteristic matrix to obtain a superposed characteristic matrix;
passing the topological matrix through a second convolutional neural network serving as a feature extractor to obtain a topological feature matrix;
multiplying the superposition characteristic matrix and the topological characteristic matrix by a matrix, and mapping the high-dimensional topological information of the topological characteristic matrix to a high-dimensional characteristic space of the superposition characteristic matrix to obtain an influence characteristic matrix based on space topology;
enabling the real-time driving voltage and the real-time temperature value of the luminous point to be adjusted to pass through a context encoder comprising an embedded layer so as to obtain a current state vector;
multiplying the current state vector by the influence feature matrix to obtain a decoding feature vector;
carrying out characteristic distribution correction on the influence characteristic matrix to obtain an enhanced generalization expression vector;
multiplying the enhanced generalization expression vector and the decoding characteristic vector according to position points to obtain an enhanced decoding characteristic vector; and
and performing decoding regression on the enhanced decoding characteristic vector through a decoder to obtain a decoding value, wherein the decoding value is a compensation voltage value.
In the above driving control method of the driving circuit, passing the driving voltage input matrix and the temperature input matrix through a first convolution neural network as a feature extractor to obtain a driving voltage feature matrix and a temperature feature matrix, respectively, includes: each layer of the first convolutional neural network respectively performs in the forward direction transmission of the layer: performing convolution processing on input data to obtain a convolution characteristic diagram; performing mean pooling based on local channel dimensions on the convolution feature map to obtain a pooled feature map; and performing nonlinear activation on the pooled feature map to obtain an activated feature map; wherein, the output of the last layer of the first convolution neural network is the driving voltage characteristic matrix and the temperature characteristic matrix, and the input of the first layer of the first convolution neural network is the driving voltage input matrix and the temperature input matrix.
In the above driving control method of the driving circuit, fusing the driving voltage characteristic matrix and the temperature characteristic matrix to obtain a superimposed characteristic matrix, including: fusing the driving voltage characteristic matrix and the temperature characteristic matrix according to the following formula to obtain the superposition characteristic matrix;
wherein the formula is:
M s ＝M 1 ⊙M 2
wherein, M s For the superimposed feature matrix, M 1 For said driving voltage characteristic matrix, M 2 Being the temperature characteristic matrix, the lines indicate dot-by-dot.
In the above drive control method of the drive circuit, passing the topological matrix through a second convolutional neural network as a feature extractor to obtain a topological feature matrix includes: each layer of the second convolutional neural network respectively performs in the forward direction transmission of the layer: performing convolution processing on input data to obtain a convolution characteristic diagram; performing mean pooling based on local channel dimensions on the convolution feature map to obtain a pooled feature map; and performing nonlinear activation on the pooled feature map to obtain an activated feature map; and the output of the last layer of the second convolutional neural network is the topological characteristic matrix, and the input of the first layer of the second convolutional neural network is the topological matrix.
In the above driving control method of the driving circuit, the step of passing the real-time driving voltage and the real-time temperature value of the light-emitting point to be adjusted through a context encoder including an embedded layer to obtain a current state vector includes: respectively converting the real-time driving voltage and the real-time temperature value of the luminous point to be regulated into input vectors by using the embedded layer of the context encoder containing the embedded layer so as to obtain a sequence of the input vectors; and globally context-based semantic encoding the sequence of input vectors using the converter of the context encoder including the embedded layer to obtain the current state vector.
In the above driving control method of the driving circuit, the performing feature distribution correction on the influence feature matrix to obtain an enhanced generalization expression vector includes: performing characteristic distribution correction on the influence characteristic matrix by using the following formula to obtain the enhanced generalization expression vector;
wherein the formula is:
wherein m is i,j Is said influenceMapping of feature matrices to [0,1]Eigenvalues within the interval, and i m i,j the representation sums the impact signature matrices along columns.
In the above driving control method of the driving circuit, the decoding and regressing the enhanced decoding feature vector by a decoder to obtain a decoded value includes: decoding the enhanced decoded feature vector using a plurality of fully-connected layers of the decoder to perform a decoding regression on the enhanced decoded feature vector to obtain the decoded value, wherein the formula is:
wherein X is the enhanced decoded feature vector, Y is the decoded value, W is a weight matrix, B is a bias vector,
representing the matrix multiplication, h (-) is the activation function.
Compared with the prior art, the driving circuit and the driving control method thereof provided by the application respectively carry out implicit association feature mining on the real-time driving voltage and the real-time temperature value of all the luminous points based on the deep neural network model, and introduce topological feature information in order to pay attention to the spatial association among the luminous points, so that the compensation voltage of each luminous point can be determined based on the global feature information of the luminous points, the problem of brightness attenuation of a display picture is improved through the compensation voltage, and the AMLED can display a high-quality picture. In this way, the panel light emission consistency and stability of the display can be optimized.
Drawings
The above and other objects, features and advantages of the present application will become more apparent by describing in more detail embodiments of the present application with reference to the attached drawings. The accompanying drawings are included to provide a further understanding of the embodiments of the application and are incorporated in and constitute a part of this specification, illustrate embodiments of the application and together with the description serve to explain the principles of the application. In the drawings, like reference numbers generally indicate like parts or steps.
Fig. 1 is a schematic diagram of a driving circuit according to an embodiment of the present application.
Fig. 2 is a block diagram of a driving circuit according to an embodiment of the present application.
Fig. 3 is a flowchart of a driving control method of a driving circuit according to an embodiment of the present application.
Fig. 4 is a schematic diagram of a driving control method of a driving circuit according to an embodiment of the present application.
Detailed Description
Hereinafter, example embodiments according to the present application will be described in detail with reference to the accompanying drawings. It should be understood that the described embodiments are only some embodiments of the present application and not all embodiments of the present application, and that the present application is not limited by the example embodiments described herein.
Overview of scenes
As described above, the Active Micro-LED (AMLED) display has the advantages of high contrast, high color saturation, and high brightness, which makes it one of the next generation hot display technologies. When the conventional AMLED display is operated, the temperature of the panel increases, so that the thermal effect of the light emitting diode, that is, the light emitting efficiency and the voltage (Vled) thereof become small, and the brightness of the AMLED display is attenuated. Accordingly, it is desired to solve the problem of thermal effect of the led and provide an AMLED display with high display quality. Therefore, a driving circuit is desired.
As shown in fig. 1, during operation, the driving transistor T1 outputs a driving signal according to the voltage stored at the first node. The anode of the light emitting diode D1 is electrically connected to the driving transistor to receive the driving signal and perform driving. The driver S1 has the role of determining the compensation voltage; the compensation voltage is determined according to the voltage value of the driving signal of each light emitting diode and the real-time temperature value of each light emitting diode, wherein the input signal is the driving signal of each light emitting diode.
Accordingly, the present inventors found that in the conventional driving circuit control scheme, the driving circuit design is mostly performed by regarding each light-emitting point as an independent light-emitting point, but actually, the light-emitting points may affect each other, for example, the heat generated by the light-emitting points close to each other may affect the brightness of the light-emitting point to be adjusted, and the influence degree between the light-emitting points at different distances is different. Therefore, the present inventors have desired to determine the compensation voltage for each light emitting point based on the global of the light emitting points so that the panel light emitting uniformity and stability of the display can be optimized.
Specifically, in the technical solution of the present application, first, a real-time driving voltage and a real-time temperature value of all light emitting points in a display are obtained through a temperature sensor and a voltage sensor disposed at each light emitting point in the display. Then, a convolutional neural network model having excellent performance in the aspect of characteristic extraction of implicit correlation is used for extracting correlation characteristics between voltages and between temperatures. Specifically, the real-time driving voltages of all light-emitting points in the display are constructed into a driving voltage input matrix, the real-time temperature values of all light-emitting points in the display are constructed into a temperature input matrix, and the driving voltage input matrix and the temperature input matrix are respectively processed through a first convolution neural network serving as a feature extractor so as to respectively extract the real-time driving voltages of all light-emitting points in the display and the implicit correlation features of the real-time temperature values of all light-emitting points in the display, thereby obtaining a driving voltage feature matrix and a temperature feature matrix. In this way, the obtained driving voltage characteristic matrix with the driving voltage correlation and the temperature characteristic matrix with the temperature correlation are fused, for example, the fusion is carried out by multiplying the position points, so that the implicit correlation between the heat generated by the light-emitting point and the brightness of the light-emitting point is focused in the subsequent voltage compensation, and the superimposed characteristic matrix is obtained. It will be appreciated that this additive effect is manifested by dot multiplication, thereby improving the accuracy of the compensation, since the current driving voltage will cause the light-emitting point to continue to emit light and generate more heat.
Moreover, it is considered that the brightness of the light-emitting point to be adjusted is affected by the heat generated by the light-emitting points close to each other, and the degrees of influence are different between the light-emitting points at different distances. Therefore, in the technical solution of the present application, a topology matrix between the light-emitting point to be adjusted in the display and all other light-emitting points in the display needs to be further constructed, where a characteristic value of each position in the topology matrix is a distance between the light-emitting point to be adjusted and each other light-emitting point. Then, in order to extract topological feature information among the luminous points, feature mining is performed on the topological matrix through a second convolutional neural network serving as a feature extractor, so that a topological feature matrix is obtained.
Thus, the high-dimensional topological information of the topological feature matrix can be mapped to the high-dimensional feature space of the superimposed feature matrix by matrix multiplication of the superimposed feature matrix and the topological feature matrix, so that the global associated feature information of the light-emitting point is concerned in the subsequent voltage compensation, and the influence feature matrix based on the spatial topology is obtained.
In this way, when the voltage of the light-emitting point to be adjusted is compensated, considering that there is a mutual association between the light-emitting points to be adjusted, it is necessary to perform global implicit association feature mining on the real-time driving voltage and the real-time temperature value of the light-emitting point to be adjusted, that is, the real-time driving voltage and the real-time temperature value of the light-emitting point to be adjusted are encoded by a context encoder including an embedded layer, so as to extract global-based high-dimensional semantic features between the light-emitting points to be adjusted, so as to obtain the current state vector.
Further, multiplying the current state vector and the influence feature matrix to map the current state vector into a high-dimensional feature space of the influence feature matrix to obtain a decoding feature vector. Then, the decoding regression is carried out on the decoding characteristic vector through a decoder, and the compensated voltage value can be obtained.
However, when the current state vector is used as a query vector and multiplied by an influence feature matrix based on a spatial topology to obtain a decoded feature vector, since the influence feature matrix expresses feature semantics based on a topology association structure between nodes, the advantage is that feature distribution as a whole between global nodes is expressed, and the generalization expression capability for a single node may have a defect, thereby affecting the regression accuracy of the decoded feature vector of the single node.
Therefore, preferably, the influence feature matrix is further subjected to semantic reasoning information explicit generalization, which is expressed as:
wherein m is i,j Is a mapping of the impact feature matrix to [0,1]Eigenvalues within the interval, and i m i,j the representation sums the impact signature matrices along columns.
Therefore, semantic concepts corresponding to the characteristic values are explicitly generalized from bottom to top into grouping instances (grouped instances) of the single node, and informatization reasoning of expression of the single node can be carried out from global characteristic semantic distribution through decoupling based on discrimination information of the grouping instances, so that distribution information plasticity of the characteristic distribution of the grouping instances in a high-dimensional semantic space of a global high-dimensional manifold is enhanced, an enhanced generalized expression vector V of the single node is obtained, and the accuracy of decoding regression is further improved.
Then, the decoding feature vector is optimized by point-multiplying the enhanced generalized expression vector V with the decoding feature vector. And then decoding and regressing in a decoder to obtain a compensation voltage value. Therefore, the driving circuit can improve the problem of brightness attenuation of the display picture by compensating the voltage, so that the AMLED can display a high-quality picture.
Based on this, the present application proposes a driving circuit comprising: the luminous point scanning module is used for acquiring real-time driving voltage and real-time temperature values of all luminous points in the display; the spatial topology construction module is used for constructing a topology matrix between a light-emitting point to be adjusted in the display and all other light-emitting points in the display, wherein the characteristic value of each position in the topology matrix is the distance between the light-emitting point to be adjusted and each other light-emitting point; the structuralization module is used for constructing real-time driving voltages of all light-emitting points in the display into a driving voltage input matrix and constructing real-time temperature values of all light-emitting points in the display into a temperature input matrix; the encoding module is used for enabling the driving voltage input matrix and the temperature input matrix to pass through a first convolution neural network serving as a feature extractor respectively so as to obtain a driving voltage feature matrix and a temperature feature matrix; the first fusion module is used for fusing the driving voltage characteristic matrix and the temperature characteristic matrix to obtain a superposed characteristic matrix; the spatial topological coding module is used for enabling the topological matrix to pass through a second convolutional neural network serving as a feature extractor so as to obtain a topological feature matrix; the second fusion module is used for multiplying the superposition characteristic matrix and the topological characteristic matrix by a matrix, and mapping the high-dimensional topological information of the topological characteristic matrix to a high-dimensional characteristic space of the superposition characteristic matrix to obtain an influence characteristic matrix based on space topology; the current state coding module is used for enabling the real-time driving voltage and the real-time temperature value of the luminous point to be adjusted to pass through a context coder comprising an embedded layer so as to obtain a current state vector; the mapping module is used for multiplying the current state vector by the influence characteristic matrix to obtain a decoding characteristic vector; the characteristic distribution correction module is used for carrying out characteristic distribution correction on the influence characteristic matrix to obtain an enhanced generalization expression vector; the optimization module is used for multiplying the enhanced generalization expression vector and the decoding characteristic vector according to position points to obtain an enhanced decoding characteristic vector; and the compensation voltage value generation module is used for decoding and regressing the enhanced decoding characteristic vector through a decoder to obtain a decoding value, and the decoding value is a compensation voltage value.
Having described the basic principles of the present application, various non-limiting embodiments of the present application will now be described with reference to the accompanying drawings.
Exemplary System
Fig. 2 illustrates a block diagram of a driving circuit according to an embodiment of the present application. As shown in fig. 2, a driving circuit 200 according to an embodiment of the present application includes: a light emitting point scanning module 210, configured to obtain real-time driving voltages and real-time temperature values of all light emitting points in the display; a spatial topology construction module 220, configured to construct a topology matrix between a light-emitting point to be adjusted in the display and all other light-emitting points in the display, where a feature value at each position in the topology matrix is a distance between the light-emitting point to be adjusted and each other light-emitting point; a structuring module 230, configured to construct real-time driving voltages of all light-emitting points in the display as a driving voltage input matrix and construct real-time temperature values of all light-emitting points in the display as a temperature input matrix; the encoding module 240 is configured to pass the driving voltage input matrix and the temperature input matrix through a first convolutional neural network as a feature extractor, respectively, to obtain a driving voltage feature matrix and a temperature feature matrix; a first fusion module 250, configured to fuse the driving voltage feature matrix and the temperature feature matrix to obtain a superimposed feature matrix; a spatial topological coding module 260, configured to pass the topological matrix through a second convolutional neural network as a feature extractor to obtain a topological feature matrix; a second fusion module 270, configured to perform matrix multiplication on the superimposed feature matrix and the topological feature matrix, and map high-dimensional topological information of the topological feature matrix into a high-dimensional feature space of the superimposed feature matrix to obtain an influence feature matrix based on spatial topology; a current state encoding module 280, configured to pass the real-time driving voltage and the real-time temperature value of the light emitting point to be adjusted through a context encoder including an embedded layer to obtain a current state vector; a mapping module 290, configured to multiply the current state vector with the influence feature matrix to obtain a decoding feature vector; a feature distribution correction module 300, configured to perform feature distribution correction on the influence feature matrix to obtain an enhanced generalization expression vector; an optimizing module 310, configured to multiply the enhanced generalized expression vector and the decoding feature vector by a position point to obtain an enhanced decoding feature vector; and a compensation voltage value generating module 320, configured to perform decoding regression on the enhanced decoding feature vector through a decoder to obtain a decoded value, where the decoded value is a compensation voltage value.
Specifically, in this embodiment of the present application, the light-emitting point scanning module 210 and the spatial topology constructing module 220 are configured to obtain real-time driving voltages and real-time temperature values of all light-emitting points in the display, and construct a topology matrix between a light-emitting point to be adjusted in the display and all other light-emitting points in the display, where a characteristic value of each position in the topology matrix is a distance between the light-emitting point to be adjusted and each other light-emitting point. As described above, since the conventional driving circuit control scheme mostly treats the light-emitting points as independent light-emitting points, the driving circuit design is actually performed, but the light-emitting points may affect each other, for example, the heat generated by the light-emitting points close to each other may affect the brightness of the light-emitting point to be adjusted, and the influence degree between the light-emitting points at different distances is different. Therefore, in the technical solution of the present application, it is desirable to determine the compensation voltage of each light emitting point based on the global of the light emitting points, so that the panel light emitting uniformity and stability of the display can be optimized.
That is, specifically, in the technical solution of the present application, first, a real-time driving voltage and a real-time temperature value of all light emitting points in the display are obtained through a temperature sensor and a voltage sensor disposed at each light emitting point in the display. Moreover, it is considered that the brightness of the light-emitting point to be adjusted is affected by the heat generated by the light-emitting points close to each other, and the degrees of influence are different between the light-emitting points at different distances. Therefore, it is also necessary to construct a topology matrix between the light-emitting point to be adjusted in the display and all other light-emitting points in the display, where the characteristic value of each position in the topology matrix is a distance between the light-emitting point to be adjusted and each other light-emitting point.
Specifically, in this embodiment of the present application, the structuring module 230 and the encoding module 240 are configured to construct real-time driving voltages of all light-emitting points in the display as a driving voltage input matrix and real-time temperature values of all light-emitting points in the display as a temperature input matrix, and pass the driving voltage input matrix and the temperature input matrix through a first convolution neural network as a feature extractor to obtain a driving voltage feature matrix and a temperature feature matrix, respectively. That is, in the technical solution of the present application, a convolutional neural network model having an excellent performance in terms of implicit associated feature extraction is further used to extract associated features between voltages and between temperatures. Specifically, in the technical scheme of the application, firstly, real-time driving voltages of all light-emitting points in the display are constructed into a driving voltage input matrix, real-time temperature values of all light-emitting points in the display are constructed into a temperature input matrix, and the driving voltage input matrix and the temperature input matrix are respectively processed through a first convolution neural network serving as a feature extractor, so that the real-time driving voltages of all light-emitting points in the display and implicit associated features of the real-time temperature values of all light-emitting points in the display are respectively extracted, and a driving voltage feature matrix and a temperature feature matrix are obtained.
More specifically, in an embodiment of the present application, the encoding module is further configured to: each layer of the first convolutional neural network performs in the forward pass of the layer: carrying out convolution processing on input data to obtain a convolution characteristic diagram; performing mean pooling based on local channel dimensions on the convolution feature map to obtain a pooled feature map; and performing nonlinear activation on the pooled feature map to obtain an activated feature map; wherein, the output of the last layer of the first convolution neural network is the driving voltage characteristic matrix and the temperature characteristic matrix, and the input of the first layer of the first convolution neural network is the driving voltage input matrix and the temperature input matrix.
Specifically, in the embodiment of the present application, the first fusing module 250 is configured to fuse the driving voltage feature matrix and the temperature feature matrix to obtain a superimposed feature matrix. That is, in the technical solution of the present application, after the driving voltage feature matrix and the temperature feature matrix are obtained, the driving voltage feature matrix having the driving voltage correlation and the temperature feature matrix having the temperature correlation are fused, for example, fused by multiplying by a position point, so as to focus on the implicit correlation between the heat generated by the light emitting point and the brightness of the light emitting point in the subsequent voltage compensation, thereby obtaining the superimposed feature matrix. It should be understood that since the current driving voltage causes the light emitting point to continue to emit light and generate more heat, the dot product is used to represent the superposition effect, thereby improving the accuracy of the voltage compensation.
More specifically, in an embodiment of the present application, the first fusion module is further configured to: fusing the driving voltage characteristic matrix and the temperature characteristic matrix according to the following formula to obtain the superposition characteristic matrix;
wherein the formula is:
M s ＝M 1 ⊙M 2
wherein M is s For the superimposed feature matrix, M 1 For said driving voltage characteristic matrix, M 2 As the temperature characteristic matrix, ", indicates dot-by-dot.
Specifically, in this embodiment of the present application, the spatial topology coding module 260 and the second fusing module 270 are configured to pass the topology matrix through a second convolutional neural network serving as a feature extractor to obtain a topology feature matrix, and perform matrix multiplication on the superimposed feature matrix and the topology feature matrix to map high-dimensional topology information of the topology feature matrix into a high-dimensional feature space of the superimposed feature matrix to obtain an influence feature matrix based on spatial topology. That is, in the technical solution of the present application, in order to extract the topological feature information between the light emitting points, feature mining is performed on the topological matrix through a second convolutional neural network serving as a feature extractor, so as to obtain a topological feature matrix. In this way, the high-dimensional topological information of the topological feature matrix can be mapped into the high-dimensional feature space of the superimposed feature matrix by matrix multiplication of the superimposed feature matrix and the topological feature matrix, so that the global associated feature information of the light-emitting point is focused in the subsequent voltage compensation process, and the influence feature matrix based on the spatial topology is obtained.
More specifically, in an embodiment of the present application, the spatial topology coding module is further configured to: each layer of the second convolutional neural network respectively performs in the forward direction transmission of the layer: performing convolution processing on input data to obtain a convolution characteristic diagram; performing mean pooling on the convolution feature map based on local channel dimensions to obtain a pooled feature map; and performing nonlinear activation on the pooled feature map to obtain an activated feature map; and the output of the last layer of the second convolutional neural network is the topological characteristic matrix, and the input of the first layer of the second convolutional neural network is the topological matrix.
Specifically, in this embodiment of the present application, the current state encoding module 280 and the mapping module 290 are configured to pass the real-time driving voltage and the real-time temperature value of the light emitting point to be adjusted through a context encoder including an embedded layer to obtain a current state vector, and multiply the current state vector and the influence feature matrix to obtain a decoding feature vector. That is, when compensating the voltage of the light emitting point to be adjusted, considering that there is a mutual association between the light emitting points to be adjusted, in the technical solution of the present application, it is necessary to perform global implicit association feature mining on the real-time driving voltage and the real-time temperature value of the light emitting point to be adjusted, that is, encoding the real-time driving voltage and the real-time temperature value of the light emitting point to be adjusted in a context encoder including an embedded layer, so as to extract global-based high-dimensional semantic features between the light emitting points to be adjusted, so as to obtain a current state vector. Further, the current state vector is multiplied by the influence feature matrix to map the current state vector into a high-dimensional feature space of the influence feature matrix to obtain a decoding feature vector.
More specifically, in this embodiment of the application, the current state encoding module is further configured to: respectively converting the real-time driving voltage and the real-time temperature value of the luminous point to be regulated into input vectors by using the embedded layer of the context encoder containing the embedded layer so as to obtain a sequence of the input vectors; and globally context-based semantic-coding the sequence of input vectors using the converter including the context encoder of the embedding layer to obtain the current state vector.
Specifically, in this embodiment, the feature distribution correction module 300 is configured to perform feature distribution correction on the influence feature matrix to obtain an enhanced generalization expression vector. It should be understood that after the decoded feature vector is obtained, the decoded feature vector is subjected to decoding regression in a decoder, so that a compensated voltage value can be obtained. However, when the current state vector is used as a query vector to be multiplied by the influence feature matrix based on the spatial topology to obtain the decoding feature vector, since the influence feature matrix expresses the feature semantics based on the topology association structure among the nodes, the advantage is that the feature distribution of the global nodes as a whole is expressed, and the generalization expression capability for a single node may have a defect, thereby affecting the regression accuracy of the decoding feature vector of the single node. Therefore, in the technical solution of the present application, preferably, the semantic reasoning information is further explicitly generalized for the impact feature matrix.
More specifically, in an embodiment of the present application, the feature distribution correction module is further configured to: performing characteristic distribution correction on the influence characteristic matrix by using the following formula to obtain the enhanced generalization expression vector;
wherein the formula is:
wherein m is i,j Is a mapping of the impact feature matrix to [0,1]Eigenvalues within the interval, and i m i,j the representation sums the impact signature matrices along columns. It should be appreciated that, thus, by explicitly generalizing the semantic concepts to which feature values correspond from bottom to top to grouped instances (grouped instances) for individual nodes, informative reasoning of the expression of individual nodes can be conducted from a global feature semantic distribution through decoupling based on discriminative information of grouped instances to enhance the feature of grouped instancesThe distribution information plasticity distributed in the high-dimensional semantic space of the global high-dimensional manifold obtains an enhanced generalized expression vector V for a single node, and further improves the accuracy of decoding regression.
Specifically, in this embodiment of the present application, the optimization module 310 and the compensation voltage value generation module 320 are configured to multiply the enhanced generalization expression vector and the decoding feature vector by a position point to obtain an enhanced decoding feature vector, and perform decoding regression on the enhanced decoding feature vector through a decoder to obtain a decoded value, where the decoded value is a compensation voltage value. That is, in the technical solution of the present application, the decoding eigenvector is optimized by further performing point multiplication on the enhanced generalization expression vector V and the decoding eigenvector, so as to obtain an enhanced decoding eigenvector, and then performing decoding regression in a decoder to obtain a compensation voltage value. Therefore, the driving circuit can improve the problem of brightness attenuation of the display picture by compensating the voltage, so that the AMLED can display a high-quality picture.
More specifically, in this embodiment of the application, the compensation voltage value generation module is further configured to: decoding the enhanced decoded feature vector using a plurality of fully-connected layers of the decoder to perform a decoding regression on the enhanced decoded feature vector to obtain the decoded value, wherein the formula is:
wherein X is the enhanced decoded feature vector, Y is the decoded value, W is a weight matrix, B is a bias vector,
representing the matrix multiplication, h (-) is the activation function.
In summary, the driving circuit 200 according to the embodiment of the present application is illustrated, which performs implicit association feature mining on real-time driving voltages and real-time temperature values of all light-emitting points respectively based on a deep neural network model, and introduces topological feature information to focus on spatial association between the light-emitting points, so that a compensation voltage of each light-emitting point can be determined based on global feature information of the light-emitting point, and the compensation voltage is used to improve the problem of brightness attenuation of a display screen, so that an AMLED can display a high-quality screen. In this way, the panel light emission consistency and stability of the display can be optimized.
As described above, the driving circuit 200 according to the embodiment of the present application may be implemented in various terminal devices, such as a server of a driving circuit algorithm, and the like. In one example, the driving circuit 200 according to the embodiment of the present application may be integrated into the terminal device as one software module and/or hardware module. For example, the driving circuit 200 may be a software module in an operating system of the terminal device, or may be an application developed for the terminal device; of course, the driving circuit 200 may also be one of many hardware modules of the terminal device.
Alternatively, in another example, the driving circuit 200 and the terminal device may be separate devices, and the driving circuit 200 may be connected to the terminal device through a wired and/or wireless network and transmit the interactive information according to an agreed data format.
Exemplary method
Fig. 3 illustrates a flowchart of a drive control method of the drive circuit. As shown in fig. 3, a drive control method of a drive circuit according to an embodiment of the present application includes the steps of: s110, acquiring real-time driving voltages and real-time temperature values of all luminous points in a display; s120, constructing a topology matrix between the light-emitting point to be adjusted in the display and all other light-emitting points in the display, wherein the characteristic value of each position in the topology matrix is the distance between the light-emitting point to be adjusted and each other light-emitting point; s130, constructing real-time driving voltages of all light-emitting points in the display into a driving voltage input matrix and constructing real-time temperature values of all light-emitting points in the display into a temperature input matrix; s140, respectively passing the driving voltage input matrix and the temperature input matrix through a first convolution neural network serving as a feature extractor to obtain a driving voltage feature matrix and a temperature feature matrix; s150, fusing the driving voltage characteristic matrix and the temperature characteristic matrix to obtain a superposition characteristic matrix; s160, passing the topological matrix through a second convolutional neural network serving as a feature extractor to obtain a topological feature matrix; s170, multiplying the superposition characteristic matrix and the topological characteristic matrix by a matrix, and mapping the high-dimensional topological information of the topological characteristic matrix to a high-dimensional characteristic space of the superposition characteristic matrix to obtain an influence characteristic matrix based on space topology; s180, enabling the real-time driving voltage and the real-time temperature value of the luminous point to be adjusted to pass through a context encoder comprising an embedded layer to obtain a current state vector; s190, multiplying the current state vector by the influence characteristic matrix to obtain a decoding characteristic vector; s200, performing characteristic distribution correction on the influence characteristic matrix to obtain an enhanced generalization expression vector; s210, multiplying the enhanced generalization expression vector and the decoding characteristic vector according to position points to obtain an enhanced decoding characteristic vector; and S220, performing decoding regression on the enhanced decoding characteristic vector through a decoder to obtain a decoding value, wherein the decoding value is a compensation voltage value.
Fig. 4 illustrates an architecture diagram of a driving control method of a driving circuit according to an embodiment of the present application. As shown in fig. 4, in the network architecture of the driving control method of the driving circuit, first, the obtained real-time driving voltages (e.g., P1 as illustrated in fig. 4) of all light emitting points in the display are constructed as a driving voltage input matrix (e.g., M1 as illustrated in fig. 4) and the real-time temperature values (e.g., P2 as illustrated in fig. 4) of all light emitting points in the display are constructed as a temperature input matrix (e.g., M2 as illustrated in fig. 4); then, passing the driving voltage input matrix and the temperature input matrix through a first convolution neural network (e.g., CNN1 as illustrated in fig. 4) as a feature extractor to obtain a driving voltage feature matrix (e.g., MF1 as illustrated in fig. 4) and a temperature feature matrix (e.g., MF2 as illustrated in fig. 4), respectively; then, fusing the driving voltage feature matrix and the temperature feature matrix to obtain a superimposed feature matrix (e.g., MF3 as illustrated in fig. 4); then, passing the obtained topological matrix (e.g., M as illustrated in fig. 4) through a second convolutional neural network (e.g., CMM2 as illustrated in fig. 4) as a feature extractor to obtain a topological feature matrix (e.g., MT as illustrated in fig. 4); then, matrix-multiplying the superimposed feature matrix and the topological feature matrix to map high-dimensional topological information of the topological feature matrix into a high-dimensional feature space of the superimposed feature matrix to obtain an influence feature matrix based on spatial topology (e.g., MF as illustrated in fig. 4); then, passing the real-time driving voltage and real-time temperature value (e.g., Q as illustrated in fig. 4) of the light-emitting point to be adjusted through a context encoder (e.g., E as illustrated in fig. 4) including an embedded layer to obtain a current state vector (e.g., VF1 as illustrated in fig. 4); then, multiplying the current state vector by the influence feature matrix to obtain a decoding feature vector (e.g., VF2 as illustrated in fig. 4); then, feature distribution correction is performed on the influence feature matrix to obtain an enhanced generalized expression vector (e.g., VF3 as illustrated in fig. 4); then, multiplying the enhanced generalized expression vector and the decoding feature vector by a position point to obtain an enhanced decoding feature vector (for example, VF as illustrated in fig. 4); and, finally, decoding the enhanced decoded feature vector by a decoder (e.g., D as illustrated in fig. 4) to obtain a decoded value, which is a compensation voltage value.
In summary, the driving control method of the driving circuit according to the embodiment of the present application is clarified, and implicit association feature mining is performed on real-time driving voltages and real-time temperature values of all light emitting points respectively based on a deep neural network model, and topological feature information is introduced to pay attention to spatial association between the light emitting points, so that compensation voltages of the light emitting points can be determined based on global feature information of the light emitting points, and the problem of luminance attenuation of a display screen is solved through the compensation voltages, so that an AMLED can display a high-quality screen. In this way, the panel light emission consistency and stability of the display can be optimized.
The basic principles of the present application have been described above with reference to specific embodiments, but it should be noted that advantages, effects, etc. mentioned in the present application are only examples and are not limiting, and the advantages, effects, etc. must not be considered to be possessed by various embodiments of the present application. Furthermore, the foregoing disclosure of specific details is for the purpose of illustration and description and is not intended to be limiting, since the foregoing disclosure is not intended to be exhaustive or to limit the disclosure to the precise details disclosed.
The block diagrams of devices, apparatuses, systems referred to in this application are only given as illustrative examples and are not intended to require or imply that the connections, arrangements, configurations, etc. must be made in the manner shown in the block diagrams. These devices, apparatuses, devices, systems may be connected, arranged, configured in any manner, as will be appreciated by those skilled in the art. Words such as "including," "comprising," "having," and the like are open-ended words that mean "including, but not limited to," and are used interchangeably therewith. As used herein, the words "or" and "refer to, and are used interchangeably with, the word" and/or, "unless the context clearly dictates otherwise. The word "such as" is used herein to mean, and is used interchangeably with, the phrase "such as but not limited to".
It should also be noted that in the devices, apparatuses, and methods of the present application, the components or steps may be decomposed and/or recombined. These decompositions and/or recombinations are to be considered as equivalents of the present application.
The previous description of the disclosed aspects is provided to enable any person skilled in the art to make or use the present application. Various modifications to these aspects will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other aspects without departing from the scope of the application. Thus, the present application is not intended to be limited to the aspects shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.
The foregoing description has been presented for purposes of illustration and description. Furthermore, the description is not intended to limit embodiments of the application to the form disclosed herein. While a number of example aspects and embodiments have been discussed above, those of skill in the art will recognize certain variations, modifications, alterations, additions and sub-combinations thereof.
Claims (10)
1. A driver circuit, comprising:
the luminous point scanning module is used for acquiring real-time driving voltages and real-time temperature values of all luminous points in the display;
the spatial topology construction module is used for constructing a topology matrix between a light-emitting point to be adjusted in the display and all other light-emitting points in the display, wherein the characteristic value of each position in the topology matrix is the distance between the light-emitting point to be adjusted and each other light-emitting point;
the structuralization module is used for constructing the real-time driving voltages of all the light-emitting points in the display into a driving voltage input matrix and constructing the real-time temperature values of all the light-emitting points in the display into a temperature input matrix;
the encoding module is used for enabling the driving voltage input matrix and the temperature input matrix to pass through a first convolution neural network serving as a feature extractor respectively so as to obtain a driving voltage feature matrix and a temperature feature matrix;
the first fusion module is used for fusing the driving voltage characteristic matrix and the temperature characteristic matrix to obtain a superposition characteristic matrix;
the spatial topological coding module is used for enabling the topological matrix to pass through a second convolutional neural network serving as a feature extractor so as to obtain a topological feature matrix;
the second fusion module is used for multiplying the superposition characteristic matrix and the topological characteristic matrix by a matrix, and mapping the high-dimensional topological information of the topological characteristic matrix to a high-dimensional characteristic space of the superposition characteristic matrix to obtain an influence characteristic matrix based on space topology;
the current state coding module is used for enabling the real-time driving voltage and the real-time temperature value of the luminous point to be adjusted to pass through a context coder comprising an embedded layer so as to obtain a current state vector;
the mapping module is used for multiplying the current state vector by the influence characteristic matrix to obtain a decoding characteristic vector;
the characteristic distribution correction module is used for carrying out characteristic distribution correction on the influence characteristic matrix to obtain an enhanced generalization expression vector;
the optimization module is used for multiplying the enhanced generalized expression vector and the decoding characteristic vector according to position points to obtain an enhanced decoding characteristic vector; and
and the compensation voltage value generation module is used for decoding and regressing the enhanced decoding characteristic vector through a decoder to obtain a decoding value, and the decoding value is a compensation voltage value.
2. The driving circuit of claim 1, wherein the encoding module is further configured to: each layer of the first convolutional neural network performs in the forward pass of the layer:
performing convolution processing on input data to obtain a convolution characteristic diagram;
performing mean pooling based on local channel dimensions on the convolution feature map to obtain a pooled feature map; and
performing nonlinear activation on the pooled feature map to obtain an activated feature map;
wherein, the output of the last layer of the first convolution neural network is the driving voltage characteristic matrix and the temperature characteristic matrix, and the input of the first layer of the first convolution neural network is the driving voltage input matrix and the temperature input matrix.
3. The driving circuit of claim 2, wherein the first fusing module is further configured to: fusing the driving voltage characteristic matrix and the temperature characteristic matrix according to the following formula to obtain the superposition characteristic matrix;
wherein the formula is:
M s ＝M 1 ⊙M 2
wherein M is s For the superimposed feature matrix, M 1 For said driving voltage characteristic matrix, M 2 As the temperature characteristic matrix, ", indicates dot-by-dot.
4. The driving circuit of claim 3, wherein the spatial topology encoding module is further configured to: each layer of the second convolutional neural network performs in the forward pass of the layer:
performing convolution processing on input data to obtain a convolution characteristic diagram;
performing mean pooling based on local channel dimensions on the convolution feature map to obtain a pooled feature map; and
performing nonlinear activation on the pooled feature map to obtain an activated feature map;
and the output of the last layer of the second convolutional neural network is the topological characteristic matrix, and the input of the first layer of the second convolutional neural network is the topological matrix.
5. The driving circuit of claim 4, wherein the current state encoding module is further configured to:
respectively converting the real-time driving voltage and the real-time temperature value of the luminous point to be regulated into input vectors by using the embedded layer of the context encoder containing the embedded layer so as to obtain a sequence of the input vectors; and
globally context-based semantic encoding the sequence of input vectors using the converter including the context encoder of the embedding layer to obtain the current state vector.
6. The driving circuit of claim 5, wherein the feature profile correction module is further configured to: performing characteristic distribution correction on the influence characteristic matrix according to the following formula to obtain the enhanced generalized expression vector;
wherein the formula is:
wherein m is i，j Is a mapping of the impact feature matrix to [0,1]Eigenvalues within the interval, and i m i，j the representation sums the impact signature matrices along columns.
7. The driving circuit of claim 6, wherein the compensation voltage value generation module is further configured to: decoding the enhanced decoded feature vector using a plurality of fully-connected layers of the decoder to perform a decoding regression on the enhanced decoded feature vector to obtain the decoded value, wherein the formula is:
wherein X is the enhanced decoded feature vector, Y is the decoded value, W is a weight matrix, B is a bias vector,
representing the matrix multiplication, h (-) is the activation function.
8. A drive control method of a drive circuit, comprising:
acquiring real-time driving voltages and real-time temperature values of all light-emitting points in a display;
constructing a topology matrix between a luminous point to be adjusted in the display and all other luminous points in the display, wherein the characteristic value of each position in the topology matrix is the distance between the luminous point to be adjusted and each other luminous point;
constructing real-time driving voltages of all light-emitting points in the display into a driving voltage input matrix and constructing real-time temperature values of all light-emitting points in the display into a temperature input matrix;
respectively passing the driving voltage input matrix and the temperature input matrix through a first convolution neural network serving as a feature extractor to obtain a driving voltage feature matrix and a temperature feature matrix;
fusing the driving voltage characteristic matrix and the temperature characteristic matrix to obtain a superposition characteristic matrix;
passing the topological matrix through a second convolutional neural network serving as a feature extractor to obtain a topological feature matrix;
multiplying the superposition characteristic matrix and the topological characteristic matrix by a matrix, and mapping the high-dimensional topological information of the topological characteristic matrix to a high-dimensional characteristic space of the superposition characteristic matrix to obtain an influence characteristic matrix based on space topology;
enabling the real-time driving voltage and the real-time temperature value of the luminous point to be adjusted to pass through a context encoder comprising an embedded layer so as to obtain a current state vector;
multiplying the current state vector by the influence feature matrix to obtain a decoding feature vector;
carrying out characteristic distribution correction on the influence characteristic matrix to obtain an enhanced generalization expression vector;
multiplying the enhanced generalization expression vector and the decoding characteristic vector according to position points to obtain an enhanced decoding characteristic vector; and
and performing decoding regression on the enhanced decoding characteristic vector through a decoder to obtain a decoding value, wherein the decoding value is a compensation voltage value.
9. The method according to claim 8, wherein the passing the driving voltage input matrix and the temperature input matrix through a first convolutional neural network as a feature extractor to obtain a driving voltage feature matrix and a temperature feature matrix, respectively, comprises:
each layer of the first convolutional neural network performs in the forward pass of the layer:
carrying out convolution processing on input data to obtain a convolution characteristic diagram;
performing mean pooling based on local channel dimensions on the convolution feature map to obtain a pooled feature map; and
performing nonlinear activation on the pooled feature map to obtain an activated feature map;
wherein, the output of the last layer of the first convolution neural network is the driving voltage characteristic matrix and the temperature characteristic matrix, and the input of the first layer of the first convolution neural network is the driving voltage input matrix and the temperature input matrix.
10. The driving control method of the driving circuit according to claim 9, wherein the performing the feature distribution correction on the influence feature matrix to obtain the enhanced generalization expression vector comprises:
performing characteristic distribution correction on the influence characteristic matrix according to the following formula to obtain the enhanced generalized expression vector;
wherein the formula is:
wherein m is i，j Is a mapping of the impact feature matrix to [0,1]Eigenvalues within the interval, and i m i，j the representation sums the impact signature matrices along columns.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210909527.0A CN115206230B (en) | 2022-07-29 | 2022-07-29 | Driving circuit and driving control method thereof |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210909527.0A CN115206230B (en) | 2022-07-29 | 2022-07-29 | Driving circuit and driving control method thereof |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115206230A true CN115206230A (en) | 2022-10-18 |
| CN115206230B CN115206230B (en) | 2023-04-28 |
Family
ID=83585649
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210909527.0A Active CN115206230B (en) | 2022-07-29 | 2022-07-29 | Driving circuit and driving control method thereof |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115206230B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116149397A (en) * | 2023-04-20 | 2023-05-23 | 福建天甫电子材料有限公司 | Temperature self-adaptive control system for electronic grade hydrogen peroxide storage |
| CN116247824A (en) * | 2023-03-30 | 2023-06-09 | 国网河南省电力公司安阳供电公司 | Control method and system for power equipment |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102005181A (en) * | 2010-11-19 | 2011-04-06 | 深圳市金立翔光电科技有限公司 | Standard dot matrix light source and image point correction method of LED display screen |
| CN105469740A (en) * | 2015-12-15 | 2016-04-06 | 昆山工研院新型平板显示技术中心有限公司 | Active matrix organic light emitting diode (OLED) and driving method thereof |
| CN107180612A (en) * | 2017-07-24 | 2017-09-19 | 京东方科技集团股份有限公司 | A kind of image element circuit and display panel |
| US20180005588A1 (en) * | 2016-06-29 | 2018-01-04 | Semiconductor Energy Laboratory Co., Ltd. | Electronic device, operation method of the electronic device, and moving vehicle |
| JP2018194719A (en) * | 2017-05-19 | 2018-12-06 | 株式会社半導体エネルギー研究所 | Display system and electronic apparatus |
| KR20210094691A (en) * | 2020-01-21 | 2021-07-30 | 삼성디스플레이 주식회사 | Afterimage preventing method and display device including the same |
| CN113223458A (en) * | 2021-01-25 | 2021-08-06 | 重庆京东方显示技术有限公司 | Pixel circuit, driving method thereof, display substrate and display device |
-
2022
- 2022-07-29 CN CN202210909527.0A patent/CN115206230B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102005181A (en) * | 2010-11-19 | 2011-04-06 | 深圳市金立翔光电科技有限公司 | Standard dot matrix light source and image point correction method of LED display screen |
| CN105469740A (en) * | 2015-12-15 | 2016-04-06 | 昆山工研院新型平板显示技术中心有限公司 | Active matrix organic light emitting diode (OLED) and driving method thereof |
| US20180005588A1 (en) * | 2016-06-29 | 2018-01-04 | Semiconductor Energy Laboratory Co., Ltd. | Electronic device, operation method of the electronic device, and moving vehicle |
| JP2018194719A (en) * | 2017-05-19 | 2018-12-06 | 株式会社半導体エネルギー研究所 | Display system and electronic apparatus |
| CN107180612A (en) * | 2017-07-24 | 2017-09-19 | 京东方科技集团股份有限公司 | A kind of image element circuit and display panel |
| KR20210094691A (en) * | 2020-01-21 | 2021-07-30 | 삼성디스플레이 주식회사 | Afterimage preventing method and display device including the same |
| CN113223458A (en) * | 2021-01-25 | 2021-08-06 | 重庆京东方显示技术有限公司 | Pixel circuit, driving method thereof, display substrate and display device |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116247824A (en) * | 2023-03-30 | 2023-06-09 | 国网河南省电力公司安阳供电公司 | Control method and system for power equipment |
| CN116247824B (en) * | 2023-03-30 | 2023-11-17 | 国网河南省电力公司安阳供电公司 | Control method and system for power equipment |
| CN116149397A (en) * | 2023-04-20 | 2023-05-23 | 福建天甫电子材料有限公司 | Temperature self-adaptive control system for electronic grade hydrogen peroxide storage |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115206230B (en) | 2023-04-28 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN115206230B (en) | Driving circuit and driving control method thereof | |
| CN111755078B (en) | Drug molecule attribute determination method, device and storage medium | |
| CN109558890A (en) | Zero sample image classification method of confrontation network is recycled based on adaptive weighting Hash | |
| CN115100678B (en) | Cross-modal pedestrian re-identification method based on channel recombination and attention mechanism | |
| US20190266474A1 (en) | Systems And Method For Character Sequence Recognition | |
| CN114677185A (en) | Intelligent large-screen advertisement intelligent recommendation system and recommendation method thereof | |
| CN111597276B (en) | Entity alignment method, device and equipment | |
| CN112397155A (en) | Single-step reverse synthesis method and system | |
| CN113780470B (en) | Graph classification method based on self-adaptive multichannel cross graph convolutional network | |
| CN113076957A (en) | RGB-D image saliency target detection method based on cross-modal feature fusion | |
| CN114580520A (en) | Monitoring system and monitoring method for secondary equipment of power system | |
| CN113111543A (en) | Internet of things service system | |
| CN113778718A (en) | Micro-service resource management method and system based on dynamic routing and electronic equipment | |
| CN117315293A (en) | Transformer-based space-time context target tracking method and system | |
| US20240232477A1 (en) | Method for predicting performance of light-emitting diode structure | |
| CN112766412B (en) | Multi-view clustering method based on self-adaptive sparse graph learning | |
| CN115617636B (en) | Distributed performance test system | |
| CN117474498A (en) | Automatic reminding system and method for patent annual fee | |
| CN110647917B (en) | Model multiplexing method and system | |
| CN116844711A (en) | Disease auxiliary identification method and device based on deep learning | |
| CN115269980A (en) | Intelligent cabinet control system based on big data and control method thereof | |
| CN116909641A (en) | Efficient and procedural GPU (graphics processing unit) expansion module group automatic configuration method | |
| CN115206754A (en) | X-ray tube and current control method thereof | |
| CN110795591B (en) | Image retrieval method based on discrete gradient back propagation | |
| CN114170459A (en) | Dynamic building identification guiding method and system based on Internet of things and electronic equipment |
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 |