Mura compensation device

Abstract

Disclosed is a mura compensation device. The mura compensation device comprises: a detection unit that detects mura on a display panel from an image of the display panel and calculates a mura compensation value for compensating for the detected mura; and an encoder that calculates residuals between the calculated mura compensation value and a curved surface predicted by applying the calculated mura compensation value to a curved surface prediction model, detects abnormal residuals and normal residuals from the calculated residuals, and compresses the detected abnormal residuals and normal residuals into compensation values of preset sizes through different codebooks.

Description

Mura compensation device

[0001] The embodiment relates to a mura compensation device.

[0002] As the information society develops, the demand for display devices for displaying images is increasing in various forms, and recently, various types of display devices such as liquid crystal display devices (LCDs) and organic light emitting display devices (OLEDs) are being utilized.

[0003] The manufacturing process for display devices involves not only assembly but also an inspection process to check image quality. This inspection process includes checking for brightness and color deviations in the display panel. When displaying an image by applying image data of the same size to the pixels of a display panel, the brightness and color should be consistent across the entire display panel. However, due to process errors, even when displaying an image using the same image data, mura defects such as smudges and color deviations can occur.

[0004] These Mura defects have a global tendency, but intermittent abnormal data that deviates from this trend can occur for various reasons. Restoration performance for these intermittent abnormal data is poor, which can result in degraded image quality. Therefore, various methods for handling abnormal data are required.

[0005] Embodiments may provide a mura compensation device capable of processing abnormal data.

[0006] The tasks of the present invention are not limited to the tasks mentioned above, and other tasks not mentioned will be clearly understood by those skilled in the art from the description below.

[0007] A mura compensation device according to an embodiment may include a detection unit that detects mura of a display panel from an image taken of the display panel and calculates a mura compensation value for compensating for the detected mura; and an encoder that applies the calculated mura compensation value and the calculated mura compensation value to a curved surface prediction model to calculate a residual from a predicted curved surface, detects an abnormal residual and a normal residual from the calculated residual, and compresses the detected abnormal residual and the normal residual into a compensation value of a predetermined size through different codebooks, respectively.

[0008] The encoder may include an abnormal data detection unit that detects abnormal residuals and normal residuals from the generated residuals; a first vector quantization unit that outputs the detected normal residuals as a compressed compensation value of a predetermined size using a first codebook generated in advance; and a second vector quantization unit that outputs the detected abnormal residuals as a compressed compensation value of a predetermined size using a second codebook generated in advance.

[0009] The above abnormal data detection unit can vectorize the produced residuals into blocks, compare each vectorized residual with a predetermined threshold, and detect a residual smaller than the threshold as a result of the comparison as the abnormal residual.

[0010] The first vector quantization unit may vectorize the produced residue in block units, sample a predetermined number of residues from the vectorized residue, compare the vector of the sampled residue with a predetermined center, assign an index identical to a cluster to which the closest center belongs as a result of the comparison, and generate a first codebook using the vector assigned with the index as a codeword.

[0011] The second vector quantization unit can vectorize the produced residual in units of blocks, compare the vector of the block unit with a predetermined center, assign an index identical to the cluster to which the closest center belongs as a result of the comparison, and generate a second codebook using the vector assigned with the index as a codeword.

[0012] The above encoder samples a predetermined number of the stored Mura compensation values, and can predict a surface for the Mura compensation values ​​based on the sampled Mura compensation values.

[0013] The encoder normalizes the coordinates of the sampled Mura compensation value, applies the normalized coordinates to the sampled Mura compensation value to calculate coefficients of the surface prediction model, and can predict the surface using the calculated coefficients and the normalized coordinates.

[0014] A mura compensation device according to an embodiment includes a memory for storing a compensation value compressed to a predetermined size for compensating for mura of a display panel; and a decoder for applying the compressed compensation value to a surface prediction model to restore a pre-predicted surface, restoring a residual from a codebook using the compressed compensation value, and restoring the mura compensation value using the restored surface and the residual, wherein the codebook may include a first codebook for restoring an abnormal residual among the restored residuals and a second codebook for restoring a normal residual among the restored residuals.

[0015] The decoder can select one codeword from the first and second codebooks using the compressed compensation value, and restore the selected codeword as a residual of the predicted surface and the Mura compensation value.

[0016] The above decoder can restore the Mura compensation value by adding the restored surface and the residual.

[0017] The present invention can effectively restore abnormal data by applying a mura compensation value and a compensation value sampled from the mura compensation value to a surface prediction model to calculate a residual of a predicted surface, detecting an abnormal residual and a normal residual from the calculated residual, and compressing the mura compensation value by vector quantizing the detected abnormal residual and the normal residual using different codebooks.

[0018] The present invention generates a codebook separately for each of abnormal data and normal data, but since the overall size of the codebook is the same, no additional memory is required.

[0019] The effects of the present invention are not limited to the effects mentioned above, and other effects not mentioned will be clearly understood by those skilled in the art from the description of the claims.

[0020] FIG. 1 is a drawing showing a display device according to an embodiment of the present invention.

[0021] Fig. 2 is a drawing for explaining the mura compensation principle of a display device according to an embodiment.

[0022] Fig. 3 is a drawing showing the detailed configuration of the Mura detection unit illustrated in Fig. 2.

[0023] Figure 4 is a drawing showing a detailed configuration of the encoder illustrated in Figure 3.

[0024] FIG. 5 is a diagram illustrating a method for compressing a Mura compensation value according to an embodiment of the present invention.

[0025] Figures 6a to 6d and Figures 7 to 12 are drawings for explaining the Mura compensation value compression process.

[0026] Fig. 13 is a drawing showing the detailed configuration of the Mura compensation unit illustrated in Fig. 2.

[0027] Figures 14 and 15 are drawings showing a process of restoring a Mura compensation value according to an embodiment of the present invention.

[0028] Figures 16a to 16c are drawings for comparing and explaining the restoration performance of the embodiment.

[0029] Hereinafter, embodiments disclosed in this specification will be described in detail with reference to the attached drawings. Regardless of the drawing numbers, identical or similar components will be given the same reference numbers, and redundant descriptions thereof will be omitted. The suffixes "module" and "part" used for components in the following description are assigned or used interchangeably only for the convenience of writing the specification, and do not in themselves have distinct meanings or roles. In addition, when describing the embodiments disclosed in this specification, if it is determined that a specific description of a related known technology may obscure the gist of the embodiments disclosed in this specification, a detailed description thereof will be omitted. In addition, the attached drawings are only intended to facilitate easy understanding of the embodiments disclosed in this specification, and the technical ideas disclosed in this specification are not limited by the attached drawings, and should be understood to include all modifications, equivalents, and substitutes included in the spirit and technical scope of the present invention.

[0030] Terms that include ordinal numbers, such as first, second, etc., may be used to describe various components, but the components are not limited by these terms. These terms are used solely to distinguish one component from another.

[0031] When a component is referred to as being "connected" or "connected" to another component, it should be understood that it may be directly connected or connected to that other component, but that there may be other components intervening. Conversely, when a component is referred to as being "directly connected" or "connected" to another component, it should be understood that there are no other components intervening.

[0032] Singular expressions include plural expressions unless the context clearly indicates otherwise.

[0033] In this specification, terms such as “include” or “have” are intended to specify the presence of a feature, number, step, operation, component, part or combination thereof described in the specification, but should be understood not to exclude in advance the possibility of the presence or addition of one or more other features, numbers, steps, operations, components, parts or combinations thereof.

[0034] In the embodiment, a new method is proposed to apply a mura compensation value and a compensation value sampled from the mura compensation value to a surface prediction model to produce a residual of a predicted surface, detect an abnormal residual and a normal residual from the produced residual, and compress the mura compensation value by vector quantizing the detected abnormal residual and normal residual using different codebooks.

[0035] FIG. 1 is a drawing showing a display device according to an embodiment of the present invention.

[0036] Referring to FIG. 1, a display device (100) according to an embodiment of the present invention may include a display panel (110) and a display driving circuit for driving the display panel. The display driving circuit may include a gate driving unit (120), a data driving unit (130), and a timing controller (140). The display device may further include a host system (150) that supplies various timing signals to the timing controller (140).

[0037] The display panel (110) may include a plurality of gate lines (G1 to Gn) and a plurality of data lines (D1 to Dm) that are arranged in a cross-section to define a plurality of pixel areas, and pixels (P) provided in each of the plurality of pixel areas.

[0038] The gate driver (120) may be disposed on one side of the display panel (110), for example, on the left side, as shown, but may also be disposed on both one side and the other side of the display panel (110), for example, on both the left and right sides, facing each other, as needed. The gate driver (120) may include a plurality of gate driver ICs (Gate Driver Integrated Circuits, not shown).

[0039] The gate driver (120) may be formed in the form of a tape carrier package in which a gate driver IC is mounted, but is not necessarily limited thereto, and the gate driver IC may be mounted directly on the display panel (110).

[0040] The data driving unit (130) converts a digital image signal transmitted from the timing controller (140) into an analog source signal and outputs it to the display panel (110). Specifically, the data driving unit (130) outputs an analog source signal to the data lines (D1 to Dm) in response to a data control signal (DCS: Data Control Signal) transmitted from the timing controller (140).

[0041] The data driving unit (130) may be disposed on one side of the display panel (110), for example, on the upper side, but may also be disposed on both one side and the other side of the display panel (110), for example, on both the upper and lower sides, facing each other, depending on the case. In addition, the data driving unit (130) may be formed in the form of a tape carrier package in which a source driver IC is mounted, but is not necessarily limited thereto.

[0042] The timing controller (140) can receive various timing signals including a vertical synchronization signal (Vsync), a horizontal synchronization signal (Hsync), a data enable (DE) signal, a clock signal (CLK), etc. from the host system (150) and generate a data control signal (DCS) for controlling the data driver (130) and a gate control signal (GCS) for controlling the gate driver (120). In addition, the timing controller (140) can receive image data (RGB) from the host system (150) and convert it into image data (RGB') in a form that can be processed by the data driver (130) and output it.

[0043] The timing controller (140) may include a Mura compensation unit. The Mura compensation unit may include a memory and a decoder. The memory stores a compressed Mura compensation value, and the decoder can decode the Mura compensation value stored in the memory to restore the Mura compensation value.

[0044] The timing controller (140) can correct image data using the restored mura compensation value. The timing controller (140) can determine a mura compensation value for each of a plurality of pixels based on the image data for each frame, and correct the image data using the determined mura compensation value.

[0045] The embodiment illustrates an example where the Mura compensation unit is implemented in the timing controller, but is not necessarily limited to this. For example, the Mura compensation unit may be implemented separately from the timing controller, or may be implemented in the host system or data driver.

[0046] The data control signal (DCS) may include a source start pulse (SSP), a source sampling clock (SSC), and a source output enable signal (SOE), and the gate control signal (GCS) may include a gate start pulse (GSP), a gate shift clock (GSC), and a gate output enable signal (GOE).

[0047] The host system (150) may be implemented as any one of a navigation system, a set-top box, a DVD player, a Blu-ray player, a personal computer (PC), a home theater system, a broadcast receiver, and a phone system.

[0048] The host system (150) can include a system on chip (SoC) with a built-in scaler to convert digital image data (RGB) of an input image into a format suitable for display on a display panel (110). The host system (150) can transmit digital image data (RGB) and timing signals to a timing controller (140).

[0049] Fig. 2 is a drawing for explaining the mura compensation principle of a display device according to an embodiment.

[0050] Referring to FIG. 2, the mura compensation device (200) according to the embodiment may include a mura detection unit (210) and a mura compensation unit (220). The mura detection unit (210) may be configured separately from the mura compensation unit (220) and may not be implemented in the display device, but is not necessarily limited thereto.

[0051] The mura detection unit (210) can detect the mura characteristic or mura of the display panel. The mura detection unit (210) can detect mura through a test image displayed on the display panel (110).

[0052] The mura detection unit (210) can calculate a mura compensation value to compensate for the detected mura, compress the calculated mura compensation value, and provide the compressed mura compensation value to the mura compensation unit. The compressed mura compensation value can be stored in the internal memory of the mura compensation unit.

[0053] The Mura compensation unit (220) applies the calculated Mura compensation value and the compensation value sampled from the Mura compensation value to a surface prediction model to calculate a residual with respect to the predicted surface, detects abnormal data (or abnormal residual) from the calculated residual, and vectorizes the detected abnormal data and normal data to compress the Mura compensation value through vector quantization.

[0054] The Mura compensation unit (220) can select a codeword from a pre-generated codebook based on vectorized abnormal data and normal data, and output the index of the selected codeword as a compressed compensation value. At this time, the Mura compensation unit (200) can generate multiple codebooks based on the vectorized abnormal data and normal data. For example, the codebooks can include a first codebook generated based on normal data and a second codebook generated based on abnormal data.

[0055] In the embodiment, since multiple Mura compensation values ​​are compressed through vector quantization, the compression efficiency can be greatly improved compared to the existing compression method through scalar quantization.

[0056] FIG. 3 is a drawing showing a detailed configuration of a mura detection unit shown in FIG. 2, FIG. 4 is a drawing showing a detailed configuration of an encoder shown in FIG. 3, FIG. 5 is a drawing showing a method for compressing a mura compensation value according to an embodiment of the present invention, and FIGS. 6a to 6d and FIGS. 7 to 12 are drawings for explaining a mura compensation value compression process.

[0057] Referring to FIGS. 3 to 12, the mura detection unit (210) of the embodiment may include an image supply unit (211), an image acquisition unit (212), a detection unit (213), a memory (214), and an encoder (215).

[0058] The image supply unit (211) can supply a test image for each predetermined grayscale to the display panel. Here, the test image for each grayscale may be a test image for each predetermined grayscale, for example, a test image for 36 grayscale, 64 grayscale, or 128 grayscale, but is not limited thereto.

[0059] The image acquisition unit (212) can acquire an image by capturing a test image for each gradation displayed on the display panel.

[0060] The detection unit (213) can detect mura of the display panel from the acquired image. The detection unit (213) can compare the acquired image with a predetermined reference image and detect the difference value between the luminance value with a defect and the luminance value without a defect as mura based on the comparison result.

[0061] The detection unit (213) can calculate a noise compensation value from a predetermined compensation formula for compensating for the detected difference value (S110). The compensation formula is defined as in the following mathematical expression 1.

[0062] [Mathematical Formula 1]

[0063] z(x,y) = a(x,y)x 2 + b(x,y)x + c(x,y)

[0064] Here, z represents the difference between the luminance values ​​with and without defects, (x, y) represents two-dimensional coordinate values, and a, b, and c represent coefficients. Mathematical expression 1 defined as a second-order compensation equation is only an example for explaining the embodiment and is not necessarily limited thereto. For example, not only a first-order or third-order or higher-order mathematical expression can be used as the compensation equation, but an offset value can also be applied.

[0065] The coefficient values ​​a, b, and c of the compensation formula for compensating for the detected difference are calculated as the Mura compensation value. The Mura compensation value can be data in the form of a matrix calculated for each pixel coordinate.

[0066] The detection unit (213) can store the Mura compensation value D(x,y) calculated for each pixel coordinate value (x, y) in the memory (214).

[0067] The Mura compensation value D(x,y) calculated in this way can be expressed on the entire screen as shown in Fig. 6a.

[0068] The encoder (215) can compress the Mura compensation value stored in the memory (214) at a predetermined compression ratio and generate a compressed Mura compensation value. Referring to FIG. 4, the encoder (215) can include a surface prediction unit (215a), a subtractor (215b), and a vector quantization unit (215c).

[0069] The surface prediction unit (215a) samples a predetermined number of stored Mura compensation values ​​D(x,y) and can predict a surface for the Mura compensation value based on the sampled Mura compensation values ​​(S120). The surface for the Mura compensation value can be modeled as a high-order polynomial as in the following mathematical expression 2.

[0070] [Equation 2]

[0071] S(x,y) = p 00 + p 10 x + p 10 y + P 20 x 2+ p 11 xy + P 02 y 2

[0072] Here, p ij (i,j=0, 1, 2, ...,n) can be coefficients.

[0073] Higher-order polynomials, such as the mathematical expression 2 above, are merely examples and are not necessarily limited thereto. Surface prediction for such Mura compensation values ​​may mean globally predicting the entire surface using sampled Mura compensation values, which are two-dimensional data. The surface S(x,y) predicted using the higher-order polynomial can be expressed as shown in Figure 6b.

[0074] For example, referring to FIG. 7, the surface prediction unit (215a) can sample a predetermined number of Mura compensation values ​​among the stored Mura compensation values ​​D(x,y) (S121). At this time, the surface prediction unit (215a) can randomly sample a predetermined number of n Mura compensation values ​​among the stored Mura compensation values.

[0075] The surface prediction unit (215a) can normalize the x and y coordinates of the n sampled Mura compensation values ​​(S122). This normalization of the coordinates may be normalization to correspond to the coordinates of the surface to be predicted.

[0076] For example, the mathematical formula for normalizing x, y coordinates can be defined as follows.

[0077] [Equation 3]

[0078] Xn = (X-Xmu) / Xstd

[0079] Yn = (Y-Ymu) / Ystd

[0080] Here, (X-Xmu) and (Y-Ymu) are the means of each x and y, and Xstd and Ystd represent the normal distributions of each x and y.

[0081] The surface prediction unit (215a) can calculate the variables of a polynomial of a predetermined degree by applying normalized coordinates to the n sampled Mura compensation values ​​(S123). For example, the variables of the polynomial according to the above mathematical expression 2 are x, y, x 2 , xy, y 2 may include.

[0082] For example, in Equation 2, which represents a matrix for n sampled Mura compensation values, the variables are {x0, y0, x0 2 , x0y0, y0 2}, {x1, y1, x1 2 , x1y1, y1 2}, ..., {x n , y n , x n 2 , x n y n , y n 2} may be.

[0083] The surface prediction unit (215a) can calculate the coefficient values ​​of a polynomial using the calculated variables (S130). The surface prediction unit (215a) calculates the coefficient values ​​of a higher-order polynomial using mathematical expression 2 (S124), and can globally predict the surface using the calculated coefficient values ​​and normalized x, y coordinates (S125).

[0084] Additionally, the surface prediction unit (215a) can predict a surface for a Mura compensation value based on the entire Mura compensation value without sampling a predetermined number of stored Mura compensation values. If sampling is not performed, the surface can be accurately predicted, but the amount of computation may increase. In the embodiment, the surface is predicted by sampling the Mura compensation value.

[0085] The subtractor (215b) can calculate a residual using the Mura compensation value and the predicted surface (S140). The subtractor (215b) calculates the residual by subtracting the predicted surface value from the Mura compensation value. The residual R(x,y) calculated in this way is defined as in the following mathematical equation 4.

[0086] [Equation 4]

[0087] R(x,y) = D(x,y) - S(x,y)

[0088] The residual R(x,y) obtained by subtracting the predicted surface value from the Mura compensation value can be expressed as shown in Fig. 6c. The more accurate the surface prediction, the more randomly distributed the residual has a narrow value range. In the embodiment, this surface prediction can suppress image quality deterioration, such as edge damage, during the vector quantization process for vectorizing and compressing the residual.

[0089] By using these Mura compensation values ​​and the predicted surface to obtain the residual, the high-frequency components can be effectively removed while preserving the low-frequency components of the data.

[0090] The vector quantization unit (215c) can generate a codebook using the produced residual. The vector quantization unit (215c) can include an abnormal data detection unit (215c-1), a first vector quantization unit (215c-2), a second vector quantization unit (215c-3), and a combination unit (215c-4).

[0091] The vector quantization process is described in detail with reference to Fig. 8 as follows.

[0092] The abnormal data detection unit (215c-1) can detect abnormal residuals from the generated residuals. The abnormal data detection unit (215c-1) can vectorize the generated residuals into block units as shown in Fig. 9 (S150).

[0093] At this time, the block (B) includes a predetermined number of residuals and can be specified in units of pixel lines. The block (B) is specified in units of pixel lines so that the capacity of the line memory is reduced, and can include a vector of 1×k (k is a positive integer). In the embodiment, the block is implemented as a vector of 1×k in units of pixel lines, but this is not necessarily limited to this. For example, the block can be implemented as a vector of k×1 or k×k.

[0094] When the residual is expressed in 8 bits, the size of the vector can be expressed as 1×k×8(bit). Here, k represents the size of the vector and can also represent the compression ratio. That is, in the embodiment, the compression ratio varies depending on the size of the vector, and the larger the vector size, the higher the compression ratio.

[0095] The abnormal data detection unit (215c-1) can rearrange the vectorized residuals in block units as shown in Fig. 10. For example, the vectorized residuals can be Vector_0, Vector_1, ..., Vector_n.

[0096] The abnormal data detection unit (215c-1) can detect abnormal residuals from the rearranged residuals. For example, the abnormal data detection unit (215c-1) can compare each rearranged residual with a predetermined threshold and detect residuals smaller than the threshold as abnormal residuals.

[0097] The first vector quantization unit (215c-2) randomly samples n predetermined normal residuals from among the normal residuals, and compares the vector of the sampled normal residuals with a randomly assigned centroid as shown in Fig. 11, and finds the closest centroid based on the comparison result, and assigns the same index as the index assigned to the cluster to which the centroid belongs. In the embodiment, clusters can be classified using K-means clustering. K-means clustering is the most widely used algorithm in unsupervised learning, and divides given data into several groups with similar characteristics.

[0098] The first vector quantization unit (215c-2) can generate a first codebook (codebook1) based on an indexed vector. The first codebook can include a plurality of codewords representing vectors in block units, and each of the plurality of codewords can be assigned an index.

[0099] The second vector quantization unit (215c-3) can compare the vector of each abnormal residual detected as shown in Fig. 11 with a randomly assigned centroid, find the closest centroid based on the comparison result, and assign an index identical to the index assigned to the cluster to which the centroid belongs.

[0100] The second vector quantization unit (215c-3) can generate a second codebook (codebook2) based on the indexed vector (S160). The second codebook can include a plurality of codewords representing block-unit vectors, and each of the plurality of codewords can be assigned an index.

[0101] The combining unit (215c-4) can combine the generated first and second codebooks to generate a single codebook. In an embodiment, a codebook having n indices can be generated, and the size of the codebook can be determined based on the number of bits representing the indices. A block-unit vector can be compressed into 8 bits, which are the codebook indices, and the compression ratio can be 8:1.

[0102] In the embodiment, the size A of the codebook can be maintained without considering the abnormal residual. As shown in Fig. 12, the sum of the size B of the first codebook and the size C of the second codebook can be equal to the size A of the codebook. For example, when the size A of the codebook is 256, the size B of the first codebook can be generated as 192 and the size C of the second codebook can be generated as 64.

[0103] The combining unit (215c-4) can select a codeword from a pre-generated codebook based on the calculated residual and output an index (Codebook Index, CI) of the selected codeword (S170). The index calculated in this way may be compressed data of the Mura compensation value.

[0104] At this time, the produced index is a compressed compensation value, and together with the index, the codebook, coordinate values, and coefficient values ​​produced in the surface prediction unit can be stored in the memory of the Mura compensation unit.

[0105] FIG. 13 is a drawing showing a detailed configuration of the Mura compensation unit illustrated in FIG. 2, and FIGS. 14 and 15 are drawings showing a process of restoring the Mura compensation value according to an embodiment of the present invention.

[0106] In the embodiment, a case where a mura compensation unit for restoring a compressed mura compensation value is implemented inside a timing controller is described as an example, but is not necessarily limited thereto.

[0107] Referring to FIGS. 13 to 15, the Mura compensation unit (220) according to the embodiment includes a memory (221), a decoder (222), and the decoder (222) may include a surface restoration unit (222a), a vector dequantization unit (222b), and an adder (222c).

[0108] The memory (221) can store the compressed compensation value from the noise detection unit and information necessary to restore the compressed compensation value, such as coordinate values, coefficient values, and a codebook.

[0109] The surface restoration unit (222a) can restore the surface using the coordinate values ​​and coefficient values ​​calculated from the surface prediction unit (S210).

[0110] The vector dequantization unit (222b) can output the codeword of the codebook as a restored residual based on the index, which is a compressed compensation value (S220). At this time, the output residual R' is an approximated value of R in which information loss has occurred. At this time, the codebook used in the vector dequantization unit (222b) may be the same as the codebook used in the vector quantization unit (215c) of the encoder (210).

[0111] The adder (222c) can output a restored mura compensation value using the restored residual and the surface. The adder (222c) can restore the mura compensation value by adding the restored residual and the surface value (S230) and output the restored mura compensation value (S240). The restored mura compensation value D'(x,y) is expressed as in FIG. 6d and can be provided for image data compensation.

[0112] Figures 16a to 16c are drawings for comparing and explaining the restoration performance of the embodiment.

[0113] Referring to FIGS. 16a to 16c, the restoration performance of an embodiment in which a codebook for abnormal data is allocated and a comparative example in which a codebook for abnormal data is not allocated are shown.

[0114] When the original data of Fig. 16a is compressed and restored using the example, the abnormal data of the original data is restored as is, as shown in Fig. 16b, whereas when the original data is compressed and restored using the comparative example, the abnormal data of the original data is not properly restored, as shown in Fig. 16c.

[0115] Through this, it can be seen that the restoration performance of the embodiment is superior to that of the comparative example.

[0116] Although the embodiments of the present invention have been described in more detail with reference to the attached drawings, the present invention is not necessarily limited to these embodiments, and various modifications may be implemented without departing from the technical spirit of the present invention. Therefore, the embodiments disclosed in this specification are not intended to limit the technical spirit of the present invention, but to explain it, and the scope of the technical spirit of the present invention is not limited by these embodiments. Therefore, it should be understood that the embodiments described above are illustrative in all respects and not restrictive. The protection scope of the present invention should be interpreted by the claims, and all technical ideas within a scope equivalent thereto should be interpreted as being included in the scope of the rights of the present invention.

Claims

1. A detection unit that detects the mura of the display panel from an image taken of the display panel and calculates a mura compensation value for compensating for the detected mura; and A mura compensation device including an encoder which calculates a residual from a predicted surface by applying the calculated mura compensation value to a surface prediction model, detects an abnormal residual and a normal residual from the calculated residual, and compresses the detected abnormal residual and normal residual into a compensation value of a predetermined size through different codebooks, respectively.

2. In paragraph 1, The above encoder, A detection unit that detects abnormal residuals and normal residuals from the above-mentioned generated residuals; A first vector quantization unit that outputs the detected normal residual as a compressed compensation value of a predetermined size using a first codebook generated in advance; and A Mura compensation device including a second vector quantization unit that outputs the detected abnormal residual as a compressed compensation value of a predetermined size using a second codebook generated in advance.

3. In paragraph 2, The above detection unit, The above-mentioned residuals are vectorized into blocks, A Mura compensation device that compares each of the vectorized residuals with a predetermined threshold, and detects residuals smaller than the threshold as abnormal residuals as a result of the comparison.

4. In paragraph 2, The above first vector quantization unit, The above-mentioned residuals are vectorized into blocks, Sample a predetermined number of residuals from the above vectorized residuals, The vector of the sampled residuals is compared with a predetermined center, and the result of the comparison is given the same index as the cluster to which the closest center belongs. A Mura compensation device that generates a first codebook using the vector to which the above index is assigned as a codeword.

5. In paragraph 2, The second vector quantization unit, The above-mentioned residuals are vectorized into blocks, The vector of the above block unit is compared with a predetermined center, and the result of the comparison is given the same index as the cluster to which the closest center belongs. A Mura compensation device that generates a second codebook using the vector to which the above index is assigned as a codeword.

6. In paragraph 1, The above encoder, A mura compensation device that samples a predetermined number of the stored mura compensation values ​​and predicts a surface for the mura compensation values ​​based on the sampled mura compensation values.

7. In paragraph 1, The above encoder, Normalize the coordinates of the sampled Mura compensation values ​​above, The coefficients of the surface prediction model are calculated by applying the normalized coordinates to the sampled Mura compensation value, A Mura compensation device that predicts a surface using the above-described coefficients and the above-described normalized coordinates.

8. Memory for storing a compensation value compressed to a predetermined size to compensate for the distortion of the display panel; and A decoder is included that applies the compressed compensation value to a surface prediction model to restore a pre-predicted surface, restores a residual from a codebook using the compressed compensation value, and restores a Mura compensation value using the restored surface and the residual. The above codebook is, A Mura compensation device comprising a first codebook for restoring abnormal residuals among the restored residuals and a second codebook for restoring normal residuals among the restored residuals.

9. In paragraph 8, The above decoder, A Mura compensation device that selects one codeword from the first and second codebooks using the compressed compensation value, and restores the selected codeword as a residual of the predicted surface and the Mura compensation value.

10. In paragraph 9, The above decoder, A mura compensation device that restores the mura compensation value by adding the restored surface and the residual.

Images