KR20210032247A - Display apparatus and image processing method thereof - Google Patents

Abstract

Disclosed is a display device which performs image processing using an artificial intelligence learning model. According to the present disclosure, the display device comprises: an image processor formed of a plurality of nodes, wherein at least one node among the plurality of nodes performs image processing on an input image using a learning model having a variable term; a display displaying the image-processed image; and a processor varying a value of the variable term to correspond to a changed option value when the option value related to the image processing is changed, and controlling the image processor to perform the image processing with the learning model having the varied value of the variable term.

Description

Display device and its image processing method {DISPLAY APPARATUS AND IMAGE PROCESSING METHOD THEREOF}

The present disclosure relates to a display device and an image processing method thereof, and more particularly, to a display device that performs image processing using an artificial intelligence learning model, and an image processing method thereof.

With the recent development of artificial intelligence technology, artificial intelligence technology is being used in various technical fields. Artificial intelligence technology consists of machine learning (for example, deep learning) and component technologies using machine learning.

Machine learning is an algorithm technology that classifies/learns the features of input data by itself, and element technology is a technology that simulates functions such as cognition and judgment of the human brain using machine learning algorithms such as deep learning. It consists of technical fields such as understanding, reasoning/prediction, knowledge expression, and motion control.

The various fields where artificial intelligence technology is applied are as follows. Linguistic understanding is a technology that recognizes and applies/processes human language/text, and includes natural language processing, machine translation, dialogue system, question and answer, and speech recognition/synthesis. Visual understanding is a technology that recognizes and processes objects like human vision, and includes object recognition, object tracking, image search, human recognition, scene understanding, spatial understanding, and image improvement. Inference prediction is a technique that logically infers and predicts information by judging, and includes knowledge/probability-based reasoning, optimization prediction, preference-based planning, and recommendation. Knowledge expression is a technology that automatically processes human experience information into knowledge data, and includes knowledge construction (data generation/classification), knowledge management (data utilization), and the like. Motion control is a technology that controls autonomous driving of a vehicle and movement of a robot, and includes movement control (navigation, collision, travel), operation control (behavior control), and the like.

Meanwhile, among the various fields described above, a technology for image processing an input image using a deep learning algorithm has emerged in relation to image improvement. However, in the case of using a conventional deep learning algorithm, a single weight and a bias are learned according to data, labels, or training loss, so that the output result of the neural network cannot be adjusted. There was. Accordingly, a post-processing module was required separately from the neural network in order to adjust the output result. However, when a post-processing module is additionally implemented in a neural network, there is a problem in that implementation is complicated and cost increases.

The present disclosure is directed to providing a display device capable of controlling an image output result without additional post-processing and an image processing method thereof.

A display apparatus according to an embodiment of the present disclosure includes a plurality of nodes, and at least one of the plurality of nodes includes an image processor that performs image processing on an input image using a learning model having a variable term; A display displaying the image-processed image; And a control processor for controlling the image processor to change the value of the variable term to correspond to the changed option value when the option value related to the image processing is changed, and to perform image processing with a learning model having the changed variable term value. Includes;

Meanwhile, an image processing method of a display device according to an embodiment of the present disclosure includes the steps of: displaying an image; Performing image processing on the input image using a learning model composed of a plurality of nodes and at least one of the plurality of nodes has a variable term; Displaying the image-processed image; and the performing of the image processing includes: if an option value related to the image processing is changed, changing a value of the variable term to correspond to the changed option value; and Includes.

1 is a block diagram illustrating a display device according to an embodiment of the present disclosure;
2 is a diagram illustrating an image processing method of a display device according to an exemplary embodiment of the present disclosure;
3 to 5 are diagrams for explaining a learning model according to an embodiment of the present disclosure;
6A and 6B are diagrams for explaining an image processing result using a learning model according to an embodiment of the present disclosure, and
7 is a block diagram illustrating in detail a display device according to an exemplary embodiment of the present disclosure.

Hereinafter, various embodiments of the present disclosure will be described with reference to the accompanying drawings. However, this is not intended to limit the technology described in the present disclosure to a specific embodiment, it should be understood to include various modifications, equivalents, and/or alternatives of the embodiments of the present disclosure. . In connection with the description of the drawings, similar reference numerals may be used for similar elements.

In the present disclosure, expressions such as "have," "may have," "include," or "may include" are the presence of corresponding features (eg, elements such as numbers, functions, actions, or parts). And does not exclude the presence of additional features.

In the present disclosure, expressions such as "A or B," "at least one of A or/and B," or "one or more of A or/and B" may include all possible combinations of items listed together. . For example, "A or B," "at least one of A and B," or "at least one of A or B" includes (1) at least one A, (2) at least one B, Or (3) it may refer to all cases including both at least one A and at least one B.

Expressions such as "first," "second," "first," or "second," used in the present disclosure may modify various elements regardless of order and/or importance, and It is used to distinguish it from other components and does not limit the components.

Some component (eg, the first component) is “(functionally or communicatively) coupled with/to)” to another component (eg, the second component) or “ When referred to as "connected to", it should be understood that the certain component may be directly connected to the other component or may be connected through another component (eg, a third component). On the other hand, when a component (eg, a first component) is referred to as being “directly connected” or “directly connected” to another component (eg, a second component), the component and the It may be understood that no other component (eg, a third component) exists between the different components.

The expression "configured to (configured to)" used in the present disclosure is, for example, "suitable for," "having the capacity to" depending on the situation. ," "designed to," "adapted to," "made to," or "capable of." The term "configured to (or set)" may not necessarily mean only "specifically designed to" in hardware. Instead, in some situations, the expression "a device configured to" may mean that the device "can" along with other devices or parts. For example, the phrase “a subprocessor configured (or configured) to perform A, B, and C” means a dedicated processor (eg, an embedded processor) for performing the operation, or executing one or more software programs stored in a memory device. By doing so, it may mean a generic-purpose processor (eg, a CPU or an application processor) capable of performing corresponding operations.

The display device 100 in the present disclosure may be a TV, but is not limited thereto, and may be implemented as an electronic device including a display. For example, the display device 100 may be implemented as various electronic devices such as a mobile phone, a tablet PC, a digital camera, a camcorder, a notebook, a desktop, a monitor, an e-book terminal, a navigation device, and a wearable device. In addition, the display device may be a fixed type or a mobile type, and may be a digital broadcasting receiver capable of receiving digital broadcasting.

Hereinafter, the present disclosure will be described in detail with reference to the drawings.

1 is a diagram for describing a display device according to an exemplary embodiment of the present disclosure.

Referring to FIG. 1, a display device 100 according to an embodiment of the present disclosure includes an image processor 110, a display 120, and a processor 130.

The image processor 110 may process an image input to the display device 100. Specifically, the image processor 110 may perform image processing on an input image using an artificial intelligence learning model.

Here, image processing includes white balance adjustment of the input image, color adjustment, color filter array interpolation, noise reduction processing, sharpening, and image processing. It may include various processing such as image enhancement and brightness adjustment.

The learning model used by the image processor 110 may be a model learned based on an artificial intelligence algorithm. Here, the artificial intelligence algorithm may be a deep neural network (DNN), a deep convolution neural network, a residual network, or the like. This learning model may be composed of a plurality of layers, that is, hierarchically, and the layer may include a plurality of nodes. In addition, at least one node among the plurality of nodes included in the learning model may include a variable term including a control parameter. The value of the variable term included in the learning model can be changed according to the value of the control parameter included in the variable term. That is, since information on a node having a variable term can be changed according to the value of the control parameter included in the variable term, the output value of the learning model may vary according to the value of the variable term. The learning model of the present disclosure will be described in detail with reference to FIGS. 3 to 5.

As a result, the image processor 110 may perform image processing on the input image using a learning model in which at least one node among the plurality of nodes has a variable term.

The image processor 110 may generate a feature map corresponding to an image input to the display device, and generate a value of a variable term included in the learning model based on the generated feature map.

Here, the feature map may include features that can be extracted from the image, such as edge information, brightness information, color information, and white balance information of the input image. The feature map includes one feature (eg, edge information) among features included in the image, but is not limited thereto.

Meanwhile, the image processor 110 may be manufactured in the form of a dedicated hardware chip for a learning model including at least one node having a variable term. A dedicated hardware chip for a learning model according to various embodiments of the present disclosure may include a Graphic Precessing Unit (GPU).

Meanwhile, the image processor 110 is a GPU (Graphic Precessing Unit), and may be implemented as a component that is distinct from the processor 130, but is not limited thereto. Various embodiments of the present disclosure may be performed by the processor 130. For example, the processor 130 may perform image processing on an input image using a learning model.

The display 120 may display various images and images provided by the display device 100. In particular, the display 120 may display an image processed by the image processor 110.

The display 120 is a display of various types such as a liquid crystal display (LCD), an organic light emitting diode (OLED), a liquid crystal on silicon (LCoS), or digital light processing (DLP). Can be implemented. However, the present invention is not limited thereto, and of course, it may be implemented as various types of displays capable of displaying images.

The processor 130 is a component for controlling the overall operation of the display device 100. For example, the processor 130 may control a plurality of hardware or software components connected to the processor 130 by driving an operating system or an application program, and may perform various data processing and operations. The processor 130 may be a central processing unit (CPU) or a graphics-processing unit (GPU), or both. The processor 130 may be implemented with at least one general processor, a digital signal processor, an application specific integrated circuit (ASIC), a system on chip (SoC), a microcomputer (MICOM), or the like.

The processor 130 may control the display 120 to display an image. In this case, the image displayed on the display 120 may be an image received from an external device (not shown) or may be an image previously stored in the display device 100.

When an option value related to image processing is changed, the processor 130 changes the value of the variable term to correspond to the changed option value, and controls the image processor 110 to perform image processing with a learning model having the changed variable term value. can do.

The processor 130 may receive a user command for changing an option value related to image processing. Here, the options related to image processing represent various factors used to process the image, such as white balance, color, color filter arrangement, sharpening, and brightness of the image, and the option values related to image processing are white balance, color, color filter. Represents the values of various factors such as array, sharpening, and brightness.

For example, the processor 130 may select one of a plurality of image processing options and receive a user's command to change a value of the selected image processing option. Specifically, the processor 130 receives a user command for changing an image processing option value through a user input interface (not shown) or processes an image from another electronic device (eg, a remote controller) through a communication interface (not shown). You can receive user commands to change the value of an option.

When receiving a user command for changing an image processing option value, the processor 130 may change the value of the variable term of the node including the variable term included in the learning model to correspond to the changed option value. Specifically, the processor 130 may change the value of the control parameter included in the variable term to correspond to the changed option value. For example, when receiving a user's command for increasing the sharpness of an image by 10%, the processor 130 may change the value of the variable term by changing the value of the control parameter so as to correspond to the value of increasing the sharpness by 10%.

Further, the processor 130 may control the image processor 110 to perform image processing with a learning model having a variable term value.

Hereinafter, in FIG. 2, an image processing method of a display device according to an exemplary embodiment will be described.

First, an image is displayed on the display (S210). In this case, the displayed image may be an image received from an external device or may be an image previously stored in the display device.

In this case, when an option value related to image processing of the displayed image is changed, a value of at least one node having a variable term among a plurality of nodes included in the learning model may be changed to correspond to the changed option value (S220).

In this case, the learning model may be a model learned based on an artificial intelligence algorithm. Here, the artificial intelligence algorithm may be a deep neural network (DNN), a deep convolution neural network, a residual network, or the like.

The learning model may be composed of a plurality of layers, that is, hierarchically, and the layer may include a plurality of nodes. In addition, at least one node among the plurality of nodes included in the learning model may include a variable term including a control parameter, and a value of the variable term may be changed according to the value of the control parameter.

In addition, image processing may be performed on the displayed image by using a learning model having a variable term that has been changed to correspond to an option value related to image processing (S230). Here, the options related to image processing represent various factors used to process the image, such as white balance, color, color filter arrangement, sharpening, and brightness of the image, and the option values related to image processing are white balance, color, and color. Represents the values of various factors such as filter arrangement, sharpening, and brightness.

For example, when an option value related to image processing increases, the value of the control parameter included in the variable term is increased to increase the value of the variable term, and image processing may be performed with a learning model having the increased value of the variable term. Conversely, when an option value related to image processing is decreased, a value of a control parameter included in a variable term may be decreased, and image processing may be performed with a learning model having a value of the reduced variable term.

Then, the image-processed image may be displayed (S240).

3 to 5 are diagrams for explaining a learning model according to an embodiment of the present disclosure.

The learning model of the present disclosure is a neural network model and may include a plurality of layers. For example, the learning model includes an input layer, a hidden layer, and an output layer. The input layer may transmit an input signal to the hidden layer, and the output layer may generate an output of the learning model based on a signal received from the hidden layer. The hidden layer is located between the input layer and the output layer, and through learning, a signal transmitted through the input layer can be changed to a predictable value.

The plurality of layers of the learning model may include a plurality of nodes disposed in each of the plurality of layers. Specifically, each of the input layer, the hidden layer, and the output layer includes a plurality of nodes, and nodes included in the input layer and nodes included in the hidden layer may be connected to each other using a connection weight. Hidden nodes included in the hidden layer and output nodes included in the output layer may be connected to each other using a connection weight.

Meanwhile, the learning model of the present disclosure may include a plurality of hidden layers. The plurality of hidden layers may be generated by deep learning or convolution deep learning. In this case, the plurality of hidden layers may include various layers such as a convolution layer, a ReLU layer, and a pooling layer.

The learning model of the present disclosure is characterized in that a variable term is included in at least one node among a plurality of nodes included in the learning model. At least one node having a variable term may be disposed in at least one hidden layer among a plurality of hidden layers of the learning node. Specifically, at least one node having a variable term may be disposed in an arbitrary layer among a plurality of layers. For example, at least one node having a variable term may be disposed in a lower middle layer close to an output layer among a plurality of layers. Specifically, when the learning model includes n hidden layers, at least one node having a variable term will be placed in at least one layer from the n/2-th layer among the n hidden layers to the hidden layer before the output layer. I can.

On the other hand, the node including the variable term included in the learning model is a node generated based on the feature map and control parameter of the input image, and a bias operation or convolution on the convolution layer It may be a node included in the learning model through an operation.

3 is a diagram for describing a learning model including a node including a variable term generated through a bias operation.

As described above, the learning model includes a feature map 310 for an input image and a node 320 having a variable term generated based on the control parameter p. Specifically, the variable term may be generated by combining the feature map and the control parameter, and multiplying the combined value by the weight of the layer to which the variable term is added. Here, the feature map may include features that can be extracted from the image, such as edge information, brightness information, color information, white balance information, and black artifact information of the input image. The feature map includes one feature (eg, edge information) among features included in the image, but is not limited thereto.

When the node 320 having the variable term is included in the L-1th layer (hereinafter, L-1 layer), it is connected to the node having the variable term included in the L-1 layer, and at least one included in the L layer The activation function of the node 330 of can be expressed by the following equation.

은 L-1 레이어에 존재하는 복수의 노드 중 변수항을 갖는 노드를 제외한 복수의 노드의 가중합을 나타낸다. 그리고, p는 제어 파라미터의 값, E(L-1, ch_max)는 입력 영상에서 추출된 특징 맵의 값을 나타낸다. 결국, W(L-1, Ch_in, Ch_out) *(p* E(L-1, ch_max))은 특징 맵의 값과 제어 파라미터의 값을 결합한 값에 L-1레이어의 가중치를 곱한 값이다. 즉, 특징 맵의 값과 제어 파라미터의 값을 결합한 값에 L-1레이어의 가중치를 곱한 값을 L-1 레이어의 가중합에 바이어스 하여 L 레이어에 포함된 노드의 값을 구할 수 있으며, L 레이어에 포함된 노드는 변수항을 갖는 노드의 영향을 받게 된다. Here, C(L, ch _out ) is connected to a node having a variable term included in the L-1 layer, and represents the value of at least one node included in the L layer, and W(L-1, Ch _in , Ch _out ) denotes a connection weight between a node included in the L-1 layer and a node included in the L layer, and C(L-1, Ch _out) denotes a value of a node included in the L-1 layer. In other words,

Denotes a weighted sum of a plurality of nodes excluding a node having a variable term among a plurality of nodes existing in the L-1 layer. In addition, p denotes a value of a control parameter, and E(L-1, ch _max ) denotes a value of a feature map extracted from an input image. In the end, W(L-1, Ch _in , Ch _out ) *(p* E(L-1, ch _max )) is obtained by multiplying the value of the feature map and the control parameter by the weight of the L-1 layer. It's a value. In other words, the value of the node included in the L layer can be obtained by biasing the value obtained by multiplying the value of the value of the feature map and the value of the control parameter by the weight of the L-1 layer to the weighted sum of the L-1 layer. Nodes included in are affected by nodes with variable terms.

Specifically, since a plurality of nodes included in the layer of the learning model are connected to a plurality of nodes included in the next layer, the node having the variable term is in the L-1th layer (hereinafter, referred to as the L-1 layer). When included, the L layer and subsequent layers may be affected by a node having a variable term. For example, if a node having a variable term is created in the L-1 layer and the value of the variable term is increased, nodes that are directly or indirectly connected to the node having the variable term among the nodes included in the L layer and subsequent layers are nodes. The value of is increased. Therefore, the output value of the learning model may be different.

Meanwhile, as nodes having a variable term in the learning model are included in the L-1 layer, the number of nodes included in the layer after the L-1 layer may increase than the number of nodes included in the L-2 layer. For example, if the L-2 layer includes 64 nodes, and the L-1 layer includes 64 nodes and a node having one variable term, the L layer and subsequent layers may include 65 nodes. have. Meanwhile, this is only an example, and the number of nodes increased in the L layer is not necessarily limited to one.

Meanwhile, when the learning model is trained, the value of p may be set to 1, and when the learning model is used, the output value of the learning model may be controlled while changing the value of p. That is, the output value of the learning model may be changed according to the value of the control parameter p. For example, if the feature map is a feature map related to sharpness, when the value of p increases, the sharpness output value of the image increases, and when the value of p decreases, the sharpness output value of the image decreases.

Meanwhile, FIG. 4 is a diagram for explaining a learning model including a variable term generated through a convolution operation.

In the case of using the convolution operation, similarly to the case of using the bias operation of FIG. 3, a node having a variable term generated based on a feature map 430 for an input image and a control parameter p may be included.

Specifically, the activation function of at least one node 450 included in the L layer may be obtained through the following equation.

C'(ch _out )=(W(ch _out )*(p*E))

C''(ch _out )=(W'(ch _out )*(p*E))

C(L, ch)= C'(ch _out )*C(L-1, ch)+C''(ch _out )

Here, (p*E) represents a value obtained by combining the value of the control parameter (p) and the value of the feature map as described above in FIG. 3. That is, the value C'(ch _out ) of the node 410 may be generated by multiplying a value obtained by combining the value of the control parameter (p) and the value of the feature map by a weight. Likewise, the value C'(ch _out ) of the node 420 may be generated by multiplying a value obtained by combining the value of the control parameter (p) and the value of the feature map by a weight. In this case, the value of the weight (W(ch _out )) used to generate the node 410 may be different from the value of the weight (W'(ch _{out)) used to generate the node 420.}

And, based on the generated value of the node 410 and the value of the node 420, the value of the node of the L layer may be determined. Specifically, the value of the node 450 of the L layer is the value of the node 440 of the L-1 layer (C(L-1, ch)) and the value of the node 410 (C'(ch _out )). It may be generated by adding the value of the node 420 (C''(ch _{out )).}

In this way, the value of the node 410 and the value of the node 420 are generated based on the control parameter p and information E of the feature map. The value of layer L is influenced by the control parameter p.

Meanwhile, FIG. 5 is a diagram illustrating a controllable learning model by changing a value of a parameter included in a ReLu layer.

As shown in FIG. 5, the learning model of the present disclosure may include a ReLu (Rectified Linear Unit).

A general ReLu function is as shown in the following equation, which is shown in the graph 510 of FIG. 5.

However, the learning model of the present disclosure may control the output value by adjusting the scale or shift of the ReLu function.

Specifically, by resetting the ReLu function as shown in the following equation, at least one node included in the ReLu layer may include a value of the ReLu function reset as follows.

X _out = X _in * P _scale + P _shift

Here, X _out may be an output value of the ReLu layer, X _in may be an input value of the ReLu layer, P _scale may be a scale value _{of the ReLu function, and P shift} may be a shift value of the ReLu function. That is, the value of the ReLu function applied to the ReLu layer may vary according to the _{P scale} and P _shift.

_{In this case, P scale} and P _shift may be set to 1 in the learning step of the learning model, and the output values of the function including the ReLu layer may be controlled while the values _{of the P scale} and P _shift are adjusted in the testing step of the learning model. have.

6 is a diagram for explaining the progress of image processing according to an embodiment of the present disclosure.

Specifically, FIG. 6A is a diagram illustrating an image displayed on the display device when p=1, that is, in a state in which the learning model is not controlled. Referring to FIG. 6A, it can be seen that leaves of trees located on the left side of the image and lines of the building image located in the middle of the image are not clearly expressed.

Meanwhile, FIG. 6B is a diagram illustrating a result image of the image processor 110 performing image processing using a learning model in which a feature map of an input image is used as a sharpness map and a control parameter p is increased. Referring to FIG. 6B, it can be seen that lines included in the image are clearly expressed, that is, the sharpness is increased, such as a leaf of a tree located on the left side of the image and a line of a building image located in the middle of the image.

7 is a block diagram illustrating in detail a configuration of a display device according to an exemplary embodiment of the present disclosure.

Referring to FIG. 7, the display device 100 may include an image processor 110, a display 120, a processor 130, a memory 140, a communication interface 150, and an input interface 160. Among these, descriptions overlapping with the image processor 110, the display 120, and the processor 130 of FIG. 1 will be omitted.

The memory 140 is a component for storing various programs and data necessary for the operation of the display device 100. The memory 140 may be implemented as a nonvolatile memory, a volatile memory, a flash memory, a hard disk drive (HDD), a solid state drive (SSD), or the like. The memory 140 is accessed by the processor 130, and data read/write/edit/delete/update by the processor 130 may be performed. In the present disclosure, the term memory refers to a memory 140, a ROM (not shown) in the processor 130, a RAM (not shown), or a memory card (not shown) mounted in the display device 100 (for example, micro SD Card, memory stick).

The memory 140 may include an image previously stored in the display device 100 or an image received from an external device (not shown).

In addition, the memory 140 may store an artificial intelligence learning model trained to perform image processing on an image displayed on the display 120 of the display device 100. Here, in the learning model, at least one node among the plurality of nodes may change an output value according to a value of a variable term that is changed including a variable term.

The communication interface 150 is a component for the display device 100 to communicate with an external device (not shown). Through the communication interface 150, the display apparatus 100 may receive an image from an external device (not shown) or may receive a user command for selecting an option and an option value related to image processing.

The communication interface 150 may include various communication modules such as a wired communication module (not shown), a short-range wireless communication module (not shown), and a wireless communication module (not shown).

Here, the wired communication module is a module for performing communication with an external device (not shown) according to a wired communication method such as wired Ethernet. In addition, the short-range wireless communication module is a module for performing communication with an external device (not shown) located in a short distance according to a short-range wireless communication method such as Bluetooth (BT), Bluetooth Low Energy (BLE), and ZigBee method. In addition, the wireless communication module is a module that is connected to an external network according to a wireless communication protocol such as WiFi and IEEE, and communicates with an external device (not shown) and a voice recognition server (not shown). In addition, the wireless communication module is based on various mobile communication standards such as 3rd Generation (3G), 3rd Generation Partnership Project (3GPP), Long Term Evolution (LTE), LTE Advanced (LTE-A), and 5G Networks. It may further include a mobile communication module for performing communication by accessing the mobile communication network.

The input interface 160 may receive a user input for controlling the display device 100. The input interface 160 may select one of a plurality of options related to image processing and receive a user input for changing an option value for the selected option. As illustrated in FIG. 7, the input interface 160 may include a touch panel for receiving a user's voice by using a microphone, a user's hand, or a stylus pen, and a button for receiving a user manipulation. However, the example of the input interface 160 illustrated in FIG. 7 is only an embodiment, and may be implemented with other input devices (eg, a keyboard, a mouse, a motion input unit, etc.).

Meanwhile, computer instructions for performing a processing operation in the display apparatus 100 according to various embodiments of the present disclosure described above may be stored in a non-transitory computer-readable medium. I can. When the computer instructions stored in the non-transitory computer-readable medium are executed by the processor of the specific device, the above-described specific device performs the processing operation of the display apparatus 100 according to the various embodiments described above.

The non-transitory readable medium refers to a medium that stores data semi-permanently and can be read by a device, rather than a medium that stores data for a short moment, such as a register, a cache, and a memory. Specifically, the above-described various applications or programs may be provided by being stored in a non-transitory readable medium such as a CD, DVD, hard disk, Blu-ray disk, USB, memory card, ROM, or the like.

In addition, although the preferred embodiments of the present disclosure have been illustrated and described above, the present disclosure is not limited to the specific embodiments described above, and the technical field to which the disclosure belongs without departing from the gist of the disclosure claimed in the claims. In addition, various modifications can be implemented by those of ordinary skill in the art, and these modifications should not be understood individually from the technical idea or prospect of the present disclosure.

100: display device 110: image processor
120: display 130: processor

Claims (15)

In the display device,
An image processor comprising a plurality of nodes, wherein at least one node among the plurality of nodes performs image processing on an input image using a learning model having a variable term;
A display displaying the image-processed image; And
A processor that changes the value of the variable term to correspond to the changed option value when the option value related to the image processing is changed, and controls the image processor to perform image processing with a learning model having the changed variable term value; Display device comprising.
The method of claim 1,
The learning model,
Included as a plurality of layers and a plurality of nodes disposed on each of the plurality of layers,
The at least one node having the variable term is disposed on an arbitrary layer among the plurality of layers.
The method of claim 1,
The processor,
When the option value related to the image processing increases, the value of the variable term is increased by increasing the value of the control parameter included in the variable term, and an image processor is used to perform image processing with a learning model having the value of the increased variable term. Controlling, display device.
The method of claim 1,
The processor,
When the option value related to the image processing decreases, the value of the variable term is decreased by decreasing the value of the control parameter included in the variable term, and an image processor is used to perform image processing with a learning model having the value of the reduced variable term. Controlling,display device
The method of claim 1,
The image processor,
A display device that generates a feature map corresponding to the input image and generates a value of the variable term based on the generated feature map.
The method of claim 1,
The learning model includes a plurality of variable terms corresponding to each of a plurality of image processing options,
The processor,
When an option value for at least one image processing option among a plurality of image processing options is changed, a value for a variable term corresponding to the image processing option among the plurality of variable terms is changed.
The method of claim 1,
The image processing,
White balance adjustment, color adjustment, color filter array interpolation, noise reduction processing, sharpening, image enhancement of the input image , A display device including various processing such as brightness adjustment.
In the image processing method,
Displaying an image;
Performing image processing on the displayed image using a learning model comprising a plurality of nodes and at least one of the plurality of nodes having a variable term; And
Including; displaying the image-processed image;
The step of performing the image processing,
And if the option value related to the image processing is changed, changing the value of the variable term to correspond to the changed option value.
The method of claim 8,
The learning model,
Consisting of a plurality of layers and a plurality of nodes disposed on each of the plurality of layers,
The at least one node having the variable term is disposed in an arbitrary layer among the plurality of layers.
The method of claim 8,
The step of varying the value of the variable term,
When the option value related to the image processing increases, increasing the value of the control parameter included in the variable term to increase the value of the variable term and performing image processing with a learning model having the increased value of the variable term; Image processing method including.
The method of claim 8,
The step of varying the value of the variable term,
When the option value related to the image processing decreases, reducing the value of the control parameter included in the variable term to decrease the value of the variable term and performing image processing with a learning model having the value of the reduced variable term; Image processing method including.
The method of claim 8,
Generating a feature map corresponding to the displayed image; And
Generating a value of the variable term based on the generated feature map.
The method of claim 8,
The learning model includes a plurality of variable terms corresponding to each of a plurality of image processing options,
The step of varying the value of the variable term,
When an option value for at least one image processing option among the plurality of image processing options is changed, changing a value for a variable term corresponding to the image processing option among the plurality of variable terms; including, image processing Way.
The method of claim 8,
The image processing,
White balance adjustment of the displayed image, color adjustment, color filter array interpolation, noise reduction processing, sharpening, image enhancement, Image processing method, including various processing such as brightness adjustment.
A computer-readable recording medium comprising a program for executing an image processing method in a display device, comprising:
Performing image processing on the input image using a learning model composed of a plurality of nodes and at least one of the plurality of nodes has a variable term; And
Providing the image-processed image; Including,
The step of performing the image processing,
And if the option value related to the image processing is changed, changing the value of the variable term to correspond to the changed option value.

Images