KR20150037203A - Device for correcting depth map of three dimensional image and method for correcting the same - Google Patents

Abstract

The present invention relates a device for correcting a depth map of three dimensional image capable of generating a depth map with superior quality, and a method for correcting the same, the device comprising: a reliability calculation unit for calculating a reliability value for each pixel based on left-eye image data and right-eye image data; a pixel-depth calculation unit for calculating a pixel-depth value for each pixel based on the left-eye image data, the right-eye image, and the reliability value; a block-depth calculation unit which classifies entire pixels, based on spatial positions thereof, into a plurality of pixel-group blocks, calculates an average-depth value for each of pixel-group blocks based on the pixel-depth values, and calculates a block-depth value for each pixel based on the average depth values; and a fusion-depth calculation unit which calculates each of the pixel-depth values and each of the block-depth values with each value corresponding to a same pixel, sets a pixel-gain value for the pixel-depth value and a block-gain value for the block-depth value during the calculation step, based on the reliability value of the same pixel, and performs the calculation step based on the set gain values to calculate a fusion-depth value for each pixel.

Description

BACKGROUND OF THE INVENTION 1. Field of the Invention [0001] The present invention relates to a depth map correction apparatus and a correction method for a three-

The present invention relates to a depth map correcting apparatus for a three-dimensional image, and more particularly, to a depth map correcting apparatus and a correcting method for a three-dimensional image capable of generating a depth map of excellent quality.

A stereoscopic three-dimensional stereoscopic image is generated by a combination of a left-eye image and a right-eye image, which are two-dimensional images. In order to generate such a stereoscopic image, information on the stereoscopic distance difference (depth difference) between the objects included in the left eye image and the right eye image is required, and this information is contained in a depth map.

Since the image quality of the three-dimensional image is influenced by the accuracy of the depth map, it is important to eliminate errors and noise in the depth map in order to improve the image quality of the three-dimensional image.

Conventionally, however, the pixel depth value and the block depth value are always calculated at a constant rate irrespective of the reliability value of the depth map of the pixel. If the reliability is low, the artifact is more emphatically emphasized, There is a problem that noise is generated more and the quality of the 3D image generated based on such depth map is lowered.

 SUMMARY OF THE INVENTION The present invention has been conceived to solve the problems as described above, and it is an object of the present invention to provide a three-dimensional image display device capable of improving the image quality of a three-dimensional image by adjusting a ratio between a pixel depth value and a block depth value according to a reliability value, There is provided a depth map correction apparatus and method for a stereoscopic image.

According to another aspect of the present invention, there is provided an apparatus for correcting a depth map for a three-dimensional (3D) stereoscopic image, comprising: a reliability calculator for calculating a reliability value for each pixel based on left eye image data and right eye image data; A pixel depth calculation unit for calculating a pixel depth value for each pixel based on the left eye image data, the right eye image data, and the reliability value; All the pixels are classified into a plurality of pixel group blocks based on their spatial positions, an average depth value is calculated for each pixel group block based on the pixel depth values, and based on the average depth value, A block depth calculator for calculating a block depth value for the block; And calculating the pixel depth value and the block depth values corresponding to the same pixel with each other, and calculating a pixel gain value and a block depth value for the pixel depth value based on the reliability value for the pixel during the calculation operation, And a fused depth calculating unit for calculating a fused depth value for each pixel by performing the calculation based on the set gain values.

And the fusion depth calculating unit increases the pixel gain value and the block gain value as the reliability value increases.

The fusion depth value is defined by Equation 1 below; &Quot; (1) " pixel depth value * pixel gain value + block depth value * block gain value; When the block gain value is a (a is an integer between 0 and 1.0), the pixel gain value is 1-a; And the value of a is linearly decreased as the reliability value increases.

Wherein the fusing depth calculating unit maintains the pixel gain value and the block gain value at a constant value irrespective of the reliability value when the reliability value is greater than a predetermined confidence threshold value; The pixel gain value is increased and the block gain value is lowered as the reliability value is decreased from the confidence threshold value.

The fusion depth value is defined by the following equation (2); &Quot; (2) &quot; pixel depth value * pixel gain value + block depth value * block gain value; When the block gain value is a (a is an integer between 0 and 1.0), the pixel gain value is 1-a; When the confidence value is less than or equal to the confidence threshold, the a is defined by: < EMI ID = 3.0 > (3) block threshold + {(0.5-block threshold) / confidence threshold * confidence value}; The block threshold value is an integer between 0.5 and 1.0; When the confidence value is greater than the confidence threshold, the value a is set to 0.5.

Wherein the reliability calculating unit searches the left eye image data for a specific pixel from the left eye image data and finds a right eye tone value for the specific pixel and a plurality of pixels adjacent to the specific pixel from the left eye image data, Selecting a right eye tone value having a value most similar to the left eye tone value among the tone values and setting a reliability value for the specific pixel based on a difference between the selected right eye tone value and the left eye tone value; The reliability value increases as the difference between the selected right-eye tone value and the left-eye tone value becomes smaller.

The block depth calculating unit further performs an operation of correcting a block depth value of pixels adjacent to the boundary by using linear interpolation in order to reduce the gradation between pixels located at the boundary between the pixel group blocks .

According to another aspect of the present invention, there is provided a depth map correction method for a three-dimensional (3D) stereoscopic image, the method comprising: A) calculating a reliability value for each pixel based on left eye image data and right eye image data; Calculating a pixel depth value for each pixel based on the left eye image data, the right eye image data, and the reliability value; All the pixels are classified into a plurality of pixel group blocks based on their spatial positions, an average depth value is calculated for each pixel group block based on the pixel depth values, and based on the average depth value, Calculating a block depth value for the block; And calculating the pixel depth value and the block depth values corresponding to the same pixel with each other, and calculating a pixel gain value and a block depth value for the pixel depth value based on the reliability value for the pixel during the calculation operation, And a D step of calculating a fused depth value for each pixel by performing the calculation on the basis of the set gain values.

The step (D) increases the pixel gain value and decreases the block gain value as the reliability value increases.

The fusion depth value is defined by Equation 1 below; &Quot; (1) &quot; pixel depth value * pixel gain value + block depth value * block gain value; When the block gain value is a (a is an integer between 0 and 1.0), the pixel gain value is 1-a; And the value of a is linearly decreased as the reliability value increases.

Wherein the step (D) comprises the steps of: maintaining a pixel gain value and a block gain value at a constant value regardless of the reliability value when the reliability value is greater than a predetermined confidence threshold value; The pixel gain value is increased and the block gain value is lowered as the reliability value is decreased from the confidence threshold value.

The fusion depth value is defined by the following equation (2); &Quot; (2) &quot; pixel depth value * pixel gain value + block depth value * block gain value; When the block gain value is a (a is an integer between 0 and 1.0), the pixel gain value is 1-a; When the confidence value is less than or equal to the confidence threshold, the a is defined by: < EMI ID = 3.0 > (3) block threshold + {(0.5-block threshold) / confidence threshold * confidence value}; The block threshold value is an integer between 0.5 and 1.0; When the confidence value is greater than the confidence threshold, the value a is set to 0.5.

Wherein the step (A) searches the left eye image data for a specific pixel from the left eye image data, finds a right eye tone value for the specific pixel and a plurality of pixels adjacent to the specific pixel, from the left eye image data, Selecting a right eye tone value having a value most similar to the left eye tone value among the tone values and setting a reliability value for the specific pixel based on a difference between the selected right eye tone value and the left eye tone value; The reliability value increases as the difference between the selected right-eye tone value and the left-eye tone value becomes smaller.

The step C further performs an operation of correcting the block depth value of the pixels adjacent to the boundary by using linear interpolation to reduce the sum of the pixels located at the boundary between the pixel group blocks .

The apparatus and method for correcting a depth map for a three-dimensional image according to the present invention have the following effects.

The present invention adjusts the ratio between the pixel depth value and the block depth value according to the reliability value, and adds the values to calculate the fusion depth value. That is, according to the present invention, a pixel with low reliability adjusts the ratio between them such that the final depth value of the pixel (i.e., the fused depth value) includes a relatively larger block depth value having a less emphasized value than the pixel depth value On the other hand, a highly reliable pixel adjusts the ratio between them such that the final depth value of the pixel (i. E., The fused depth value) contains relatively more pixel depth values that have a more emphasized value than the block depth value. Accordingly, the present invention can generate a depth map of higher quality than the conventional method of summing the pixel depth value and the block depth value at a constant rate regardless of the reliability value. Therefore, the present invention can improve the image quality of a three-dimensional image.

1 is a block diagram of an apparatus for correcting a depth map for a three-dimensional image according to an embodiment of the present invention.
2 is a diagram for explaining a process of generating a pixel depth map DP-mp generated from the pixel depth calculating unit of FIG.
3 is a diagram for explaining a reliability value calculating method
4 is a diagram showing left eye image data and right eye image data applied to the pixel depth calculating unit and the reliability calculating unit of FIG. 1;
5 is a view showing a reliability map generated based on the left eye image data and the right eye image data of FIG. 4
6 is a diagram for explaining a process of generating a block depth map;
7 is a graph showing the relationship between the reliability value and the block gain value
8 is another graph showing the relationship between the reliability value and the block gain value
9 is a view for explaining a comparison between the conventional depth map of fusion and the depth map of the present invention
FIG. 10 is a view for explaining a comparison between a three-dimensional image generated based on a conventional fusion depth map and a three-dimensional image generated based on a fusion depth map according to the present invention

1 is a block diagram of an apparatus for correcting a depth map for a three-dimensional image according to an embodiment of the present invention.

1, the apparatus for correcting a depth map for a three-dimensional image according to an embodiment of the present invention includes a reliability calculation unit 101, a pixel depth calculation unit 103, a block depth calculation unit 105, And a depth calculating section 107.

The pixel depth calculating unit 103 receives the left eye image data Img_L and the right eye image data Img_R and the reliability map CF_mp (map showing the reliability value of each pixel) from the reliability calculating unit 101, Map (PD-mp). Here, the left eye image data Img_L and the right eye image data Img_R are two-dimensional image data for images (left eye image, right eye image) obtained by photographing the same image at different angles. The left eye image data Img_L and the right eye image data Img_R are all image data of one frame. The pixel depth map (DP-mp) is a map that expresses the relative distance between a plurality of subjects included in an image of one frame by a gradation difference.

With reference to Fig. 2, the pixel depth map PD-mp described above will be described in more detail.

FIG. 2 is a diagram for explaining a process of generating a pixel depth map DP-mp generated from the pixel depth calculating unit 103 of FIG.

As shown in FIG. 2, both the left eye image data Img_L and the right eye image data Img_R include a circle CC and a triangle TA. Assume here that the circle CC is projected further forward than the triangle TA and that the triangle TA is located considerably behind the circle CC. Generally, as the subject further protrudes further, the amount of positional change between the subject in the left-eye image data Img_L and the subject in the right eye image data Img_R further increases. On the other hand, as the subject is located further behind, the amount of positional change of the subject decreases. For example, as shown in FIG. 2, it can be seen that the position of the circle CC included in the left eye image data Img_L and the position of the circle CC included in the right eye image data Img_R are significantly different have. On the other hand, the positions of the triangle TA included in the left eye image data Img_L and the positions of the triangle TA included in the right eye image data Img_R are substantially the same.

The pixel depth calculating unit 103 compares the left eye image data Img_L and the right eye image data Img_R to compare the right eye image data Img_R with the circles CC and triangles TA included in the left eye image data Img_L, From the position where the circle (CC) and the triangle (TA) are included. Then, based on the calculated result, the gradation value of the pixels constituting the circle CC in the left eye image data Img_L and the gradation value of the pixels constituting the triangle TA are set, Map (PD-mp).

As shown in the pixel depth map PD-mp in Fig. 2, the circle CC, which is a relatively moving subject, is represented by a bright gradation value (i.e., a high pixel depth value) A triangle (TA), which is a very small subject, is represented by a dark tone value (i.e., a low pixel depth value). In other words, the more projected subjects are displayed as bright tonal values, while the darker tonal values are displayed as the darker tonal values of the subject located behind.

On the other hand, the reliability calculation unit 101 calculates the reliability value for the pixel depth value of each pixel by considering the similarity between the left eye image and the right eye image as described above on a pixel-by-pixel basis. For this, the reliability calculation unit 101 calculates a reliability value for each pixel based on the left eye image data and the right eye image data, and generates a reliability map CF_mp including a pixel-by-pixel reliability value. Specifically, the reliability calculation unit 101 searches the left eye image data for each pixel from the left eye image data, and calculates a right eye tone value for each pixel and a few peripheral pixels adjacent to the pixel from the left eye image data And selects a right eye tone value having a value most similar to the left eye tone value among the right eye tone values and sets a reliability value for each pixel based on the difference between the selected right eye tone value and the left eye tone value do. Here, the reliability value increases as the difference between the selected right-eye tone value and left-eye tone value becomes smaller. This reliability value calculating method will be described in more detail with reference to FIG.

3 is a diagram for explaining a reliability value calculating method.

The left eye image data Img_L and the right eye image data Img_R in FIGS. 3 (a) and 3 (b) are two-dimensional images of images (left eye image, right eye image) obtained by photographing the same image at different angles Data. Here, the displayed number of pixels means the tone value of the image data supplied to the pixel, and a pixel for which no number is displayed means that the tone value of the image data is zero.

Here, the image data shown in the left eye image data, for example, the image data corresponding to the 12th gradation and the image data corresponding to the 13th gradation may not be seen from the right eye image data by the angle difference.

3, for example, when calculating the reliability value for a specific pixel (hereinafter, referred to as the first pixel PXL), the reliability calculating unit 101 calculates the left- From the video data Img_L. That is, as shown in FIG. 3 (a), it can be seen that the left eye tone value of the first pixel PXL1 is 3. The reliability calculation unit 101 then searches the right eye image data Img_R for the right eye tone values for the first pixel PXL1 and a few peripheral pixels adjacent to the first pixel PXL1. 3B, the reliability calculating unit 101 sets the first block BL1 around the first pixel PXL1 and sets the first block BL1 in the first block BL1 as shown in FIG. The right eye tone values for all the pixels located are found, and the right eye tone values are 0, 2, 3, 5, 6, and 11, respectively. Next, the reliability calculation unit 101 selects one right eye tone value having the most similar value to the left eye tone value from the right eye tone values, and based on the difference between the selected right eye tone value and the left eye tone value And sets a reliability value for each pixel. That is, as shown in FIG. 3B, the reliability calculating unit 101 selects a gray level value most similar to the above-described left-eye gray level value among the right-eye gray level values in the first block BL1 As can be seen from FIG. 3 (b), this is 3. At this time, since the difference in gray level value between them is 0, the first pixel PXL1 has a fairly high reliability value. For example, when the reliability value has a value from 0 to 255, the first pixel PXL1 has the highest reliability value of 255. [

On the other hand, the right eye grayscale value of the second pixel PXL2 is 13, and the left eye grayscale values 7, 8, 9 and 12 0 in the second block BL2 set around the second pixel PXL2 The value closest to the right-eye grayscale value is 9. Therefore, the second pixel PXL2 exhibits a low reliability value, and such a second pixel PXL2 may appear in the form of noise in an actual screen.

When the reliability values are individually calculated for all the pixels in this manner, a reliability map CF_mp composed of the reliability values is created.

The above-described pixel-depth calculating unit 103 calculates the pixel-depth map PD-mp by correcting the pixel-depth value for each pixel using the above-described reliability map CF_map in the process of generating the pixel- ).

4 is a diagram showing left eye image data Img_L and right eye image data Img_R applied to the pixel depth calculating unit 103 and the reliability calculating unit 101 of FIG. The reliability map CF_mp generated based on the data Img_L and the right eye image data Img_R. Here, in FIG. 5, the reliability value of the pixels in the portion indicated by the bright color is higher, and the reliability value of the pixels in the dark portion is lower.

The block depth calculating unit 105 classifies all the pixels into n (n is a natural number greater than 1) pixel group blocks BLK based on their spatial positions, and outputs the pixel depth map PD- The average depth value for each pixel group block BLK is calculated on the basis of the pixel depth values from the pixel depth values obtained from the pixel depth values and the block depth value for each pixel is calculated based on the average depth value.

The block depth calculating unit 105 calculates an average depth value for each pixel group BLK by referring to pixel depth values of pixels included in the pixel depth map DP-mp, And a block depth map BD_mp containing the block depth map BD_mp. For example, the average depth value for a specific pixel group block BLK is calculated by adding pixel depth values of all pixels included in the specific pixel group block BLK, and adding the added value to the specific pixel group block BLK By the number of pixels included in the pixel. The block depth calculator 105 calculates a block depth value of pixels located adjacent to the boundary by using linear interpolation in order to reduce the gradation between pixels located at the boundary between the pixel group BLKs May be further performed.

FIG. 6 is a diagram for explaining a process of generating a block depth map. The pixel depth map DP-mp shown in FIG. 6A is divided into 810 (27 * 30) pixel group blocks BLK The average depth value is obtained for each pixel group block BLK, and the block depth value for each pixel is set to the average depth value of the pixel group block BLK to which the pixel belongs, thereby generating a block depth value for each pixel . Therefore, the pixels included in one pixel group block BLK have the same block depth values. However, since the block depth values of the boundary pixels are corrected by the linear interpolation after the operation of applying the average depth value, the pixels included in one pixel group BLK always have the same block depth value .

When the average depth value calculation operation and the linear interpolation method for each pixel group are performed as described above, the block depth map BD-mp shown in FIG. 6B is completed.

The fusing depth calculating unit 107 receives the pixel depth map PD-mp from the pixel depth calculating unit 103 and the block depth map BD-mp from the block depth calculating unit 105. Thereafter, the fused depth calculating unit 107 calculates the pixel depth values and the block depth values corresponding to the same pixel, calculates the fiducial depth values and the block depth values corresponding to the pixel depth values based on the reliability values of the pixels, A fuse depth value for each pixel is calculated by setting a block gain value for a pixel gain value and a block depth value and performing the calculation based on the set gain values, To generate a fusion depth map (FD-mp) that includes.

At this time, the fused depth calculating unit 107 can set the pixel gain value and the block gain value by the following two methods, which will be described with reference to FIGS. 7 and 8. FIG.

7 is a graph showing the relationship between the reliability value and the block gain value. As shown in the figure, the block gain value decreases linearly in inverse proportion to the reliability value. On the other hand, the pixel gain value is determined according to the block gain value, and the pixel gain value increases linearly with the increase of the reliability value.

Based on the graph shown in FIG. 7 and the following equation (1), the fused depth calculating unit 107 can calculate the fused depth value for each pixel as follows.

[Equation 1]

FusionDepth = PixelDepth * (1-a) + BlockDepth * a

In Equation 1, FusionDepth denotes a fused depth value of a pixel, PixelDepth denotes a pixel depth value of the pixel, BlockDepth denotes a block depth value of the pixel, a denotes a block gain value according to the reliability of the pixel, (1-a) denotes a pixel gain value according to the reliability of the pixel. Here, the block gain value a is an integer selected between 0 and 1.0.

8 is a graph showing the relationship between the reliability value and the block gain value. As shown in the figure, when the reliability value is larger than the predetermined confidence threshold value th2, the block gain value The block gain value is linearly decreased in proportion to the decrease of the reliability value from its confidence threshold value th2. On the other hand, the pixel gain value is determined according to the block gain value. When the pixel gain value is larger than the above-described reliability threshold value th2, the pixel gain value is maintained at a constant value regardless of the reliability value thereof, And increases linearly in proportion to the decrease from the value th2.

Based on the graph as shown in FIG. 8 and the following equation (2), the fusing depth calculating unit 107 can calculate the fusing depth value for each pixel as follows.

&Quot; (2) &quot;

FusionDepth = PixelDepth * (1-a) + BlockDepth * a

In Equation (2), FusionDepth denotes a fused depth value of a pixel, PixelDepth denotes a pixel depth value of the pixel, BlockDepth denotes a block depth value of the pixel, a denotes a block gain value according to the reliability of the pixel, (1-a) denotes a pixel gain value according to the reliability of the pixel.

At this time, when the reliability value is less than or equal to the trust threshold value th2, the above-mentioned block gain value a is defined by the following equation (3).

&Quot; (3) &quot;

a = th1 + {(0.5-th1) / th2 * Confidence_Level}

In Equation (3), th1 denotes a block threshold value, th2 denotes a confidence threshold value, and Confidence_Level denotes a confidence value. Here, the block threshold value th1 is an integer selected between 0.5 and 1.0.

On the other hand, when the reliability value is larger than the trust threshold value th2, the aforementioned block gain value a is maintained at a constant value, for example, 0.5.

As described above, according to the present invention, the ratio between the pixel depth value and the block depth value is adjusted according to the reliability value, and the fused depth value is calculated by summing the ratio. In other words, a pixel with a low reliability has a very high probability that its pixel depth value will directly affect the occurrence of an artifact, so that the final depth value of the pixel (i.e., the fused depth value) Adjust the ratio between them to include relatively more block depth values with less emphasized values. Conversely, a pixel with a high reliability has a very high probability that its pixel depth value is part of the original image, so that the final depth value (i.e., fused depth value) of the pixel has a pixel depth value that is more emphasized than the block depth value Adjust the ratio between them to include relatively more. Accordingly, the present invention can generate a depth map of higher quality than the conventional method of summing the pixel depth value and the block depth value at a constant rate regardless of the reliability value. Therefore, the present invention can improve the image quality of a three-dimensional image.

FIG. 9 is a diagram for explaining a comparison between a conventional depth map FD-mp and a depth map FD-mp according to the present invention.

9 (a) shows a conventional fused depth map FD-mp, and Fig. 9 (b) shows a fused depth map FD-mp according to the present invention. FD-mp) are all generated based on the same pixel depth map (PD-mp) and block depth map (BD-mp). Here, since the pixel depth map PD-mp includes a considerably large number of pixels with low reliability values, the fused depth map FD-mp is obtained at a fixed ratio irrespective of the reliability value, ), There are many portions highlighted with bright colors. As a result, they are likely to act as artifacts, and the image quality of a three-dimensional image is likely to deteriorate. However, as shown in FIG. 9 (b), when the convergence depth map FD-mp is created by increasing the ratio of the relatively smoothingly processed block depth values in consideration of the low reliability value Artifacts are significantly reduced and the image quality of the three-dimensional image can be improved.

FIG. 10 is a diagram for explaining a comparison between a 3-dimensional image generated based on a conventional fused depth map (FD-mp) and a 3-dimensional image generated based on a fused depth map (FD-mp) according to the present invention.

10 (a) is a three-dimensional image generated based on a conventional fused depth map (FD-mp), and FIG. 10 (b) is a three-dimensional image generated based on a fused depth map As a result, it can be seen that the image shown in FIG. 10 (b) is more excellent in image quality than the image shown in FIG. 10 (a).

It will be apparent to those skilled in the art that various modifications and variations can be made in the present invention without departing from the spirit or scope of the general inventive concept as defined by the appended claims and their equivalents. Will be clear to those who have knowledge of.

101: reliability calculating unit 103: pixel depth calculating unit
105: block depth calculating unit 107: fused depth calculating unit
Img_L: Left eye image data Img_R: Right eye image data
CF-mp: reliability map PD-mp: pixel depth map
BD-mp: Block depth map FD-mp: Fusion depth map

Claims (14)

A reliability calculating unit for calculating a reliability value for each pixel based on left eye image data and right eye image data;
A pixel depth calculation unit for calculating a pixel depth value for each pixel based on the left eye image data, the right eye image data, and the reliability value;
All the pixels are classified into a plurality of pixel group blocks based on their spatial positions, an average depth value is calculated for each pixel group block based on the pixel depth values, and based on the average depth value, A block depth calculator for calculating a block depth value for the block; And
The pixel depth values and the block depth values corresponding to the same pixel are calculated with respect to each other, and based on the reliability value for the pixel, And a fused depth calculating unit for calculating a fused depth value for each pixel by setting the block gain value and performing the calculation based on the set gain values. The method according to claim 1,
Wherein the fusing depth calculating unit raises the pixel gain value and decreases the block gain value as the reliability value is higher. 3. The method of claim 2,
The fusion depth value is defined by Equation 1 below;
&Quot; (1) &quot;
Pixel depth value * pixel gain value + block depth value * block gain value;
When the block gain value is a (a is an integer between 0 and 1.0), the pixel gain value is 1-a;
Wherein the value of a is linearly decreased as the reliability value increases. The method according to claim 1,
The fusion depth calculating unit
Maintaining the pixel gain value and the block gain value at a constant value regardless of the reliability value when the reliability value is greater than a predetermined confidence threshold value; And,
Wherein the pixel gain value is increased and the block gain value is decreased as the reliability value is decreased from the confidence threshold value. 5. The method of claim 4,
The fusion depth value is defined by the following equation (2);
&Quot; (2) &quot;
Pixel depth value * pixel gain value + block depth value * block gain value;
When the block gain value is a (a is an integer between 0 and 1.0), the pixel gain value is 1-a;
When the confidence value is less than or equal to the confidence threshold, the a is defined by: < EMI ID = 3.0 >
&Quot; (3) &quot;
Block threshold + {(0.5-block threshold) / confidence threshold * confidence value};
The block threshold value is an integer between 0.5 and 1.0; And,
And when the confidence value is greater than the confidence threshold, the value a is set to 0.5. The method according to claim 1,
Wherein the reliability calculating unit searches the left eye image data for a specific pixel from the left eye image data and finds a right eye tone value for the specific pixel and a plurality of pixels adjacent to the specific pixel from the left eye image data, Selecting a right eye tone value having a value most similar to the left eye tone value among the tone values and setting a reliability value for the specific pixel based on a difference between the selected right eye tone value and the left eye tone value; And,
Wherein the reliability value is increased as the difference between the selected right-eye gradation value and the left-eye gradation value is smaller. The method according to claim 1,
Wherein the block depth calculating unit comprises:
Further comprising the step of correcting a block depth value of pixels adjacent to the boundary by using linear interpolation in order to reduce a gradation between pixels located at a boundary between the pixel group blocks, Depth map correction device for stereoscopic image. A step of calculating a reliability value for each pixel based on left eye image data and right eye image data;
Calculating a pixel depth value for each pixel based on the left eye image data, the right eye image data, and the reliability value;
All the pixels are classified into a plurality of pixel group blocks based on their spatial positions, an average depth value is calculated for each pixel group block based on the pixel depth values, and based on the average depth value, Calculating a block depth value for the block; And
The pixel depth values and the block depth values corresponding to the same pixel are calculated with respect to each other, and based on the reliability value for the pixel, Calculating a fused depth value for each pixel by setting a block gain value, and performing the calculation based on the set gain values, to calculate a fused depth value for each pixel. 9. The method of claim 8,
Wherein the step (D) increases the pixel gain value and decreases the block gain value as the reliability value is higher. 10. The method of claim 9,
The fusion depth value is defined by Equation 1 below;
&Quot; (1) &quot;
Pixel depth value * pixel gain value + block depth value * block gain value;
When the block gain value is a (a is an integer between 0 and 1.0), the pixel gain value is 1-a;
And the value of a linearly decreases as the reliability value increases. 9. The method of claim 8,
The step (D)
Maintaining the pixel gain value and the block gain value at a constant value regardless of the reliability value when the reliability value is greater than a predetermined confidence threshold value; And,
Wherein the pixel gain value is increased and the block gain value is lowered as the reliability value is decreased from the confidence threshold value. 12. The method of claim 11,
The fusion depth value is defined by the following equation (2);
&Quot; (2) &quot;
Pixel depth value * pixel gain value + block depth value * block gain value;
When the block gain value is a (a is an integer between 0 and 1.0), the pixel gain value is 1-a;
When the confidence value is less than or equal to the confidence threshold, the a is defined by: < EMI ID = 3.0 >
&Quot; (3) &quot;
Block threshold + {(0.5-block threshold) / confidence threshold * confidence value};
The block threshold value is an integer between 0.5 and 1.0; And,
And when the confidence value is greater than the confidence threshold, the value a is set to 0.5. 9. The method of claim 8,
Wherein the step (A) searches the left eye image data for a specific pixel from the left eye image data, finds a right eye tone value for the specific pixel and a plurality of pixels adjacent to the specific pixel, from the left eye image data, Selecting a right eye tone value having a value most similar to the left eye tone value among the tone values and setting a reliability value for the specific pixel based on a difference between the selected right eye tone value and the left eye tone value; And,
Wherein the reliability value is increased as the difference between the selected right eye tone value and the left eye tone value is smaller. 9. The method of claim 8,
In the step C,
Further comprising the step of correcting a block depth value of pixels adjacent to the boundary by using linear interpolation in order to reduce a gradation between pixels located at a boundary between the pixel group blocks, Depth Map Correction Method for Stereoscopic Image.

Images