TW201325202A - Three-dimension image processing method - Google Patents

Abstract

A three-dimension image processing method includes: generating a three-dimension image having a first eye frame and a second eye frame from an original two-dimension image; generating first and second mask areas at first and second boundaries of the first eye frame; and generating third and fourth mask areas at first and second boundaries of the second eye frame. Each length of the first and the fourth mask areas includes a length of a comparison area whose length is determined based on a pixel data difference obtained by comparing the first eye frame and the second eye frame. Each length of the first to the fourth mask area further includes a length of a first extension border area.

Description

3D image processing method

The disclosure relates to a three-dimensional image processing method.

Nowadays, viewing 3D images has better entertainment effects, so more and more display devices (such as 3D TVs) support 3D image display. However, the image signal received by the three-dimensional display device may be only a two-dimensional image signal, and therefore, the three-dimensional display device converts the two-dimensional image into a three-dimensional image.

When converting a 2D image into a 3D image, it can also be called 3D wrapping, and must refer to a so-called depth map. The so-called "depth" refers to the degree to which the human eye feels when looking at an object in the image. The depth map is composed of a number of depth bits, each depth representing the depth of a pixel in the corresponding 2D image. In this way, a stereoscopic image can be provided to the viewer from an image of a known viewing angle and its corresponding depth map.

The 3D image includes a left eye picture and a right eye picture. When the human eye views the three-dimensional image, if the left eye image seen by the left eye and the right eye image viewed by the right eye have parallax, the human eye will feel that the object has a three-dimensional effect. Conversely, if there is no parallax, the human eye will feel that the object is a plane.

In general, if you display an object in the distance, move the left-eye picture to the left and the right-eye picture to the right. In contrast, to display the object in the near side, move the left eye picture to the right and the right eye picture to the left. After querying the depth map, you can see in which direction the left-eye and right-eye images are to be translated, and what the amount of translation is.

However, when processing a three-dimensional image, a border is generated at the boundary between the left-eye image and the right-eye image, but the frame affects the visual content of the three-dimensional image and affects the comfort of the human eye.

Embodiments disclosed herein relate to a three-dimensional image processing method that produces an asymmetrical virtual border.

The disclosed embodiments relate to a three-dimensional image processing method, and the generated virtual border and the three-dimensional image may be in different visual planes.

According to an exemplary embodiment, a method for processing a three-dimensional image includes: generating a first eye image and a second eye image according to one of the original two-dimensional images; and the first eye image a first mask region and a second mask region are respectively generated at a first boundary and a second boundary; and a first boundary is formed at a first boundary and a second boundary of the second eye A three-mask area and a fourth mask area. The length of each of the first and fourth mask regions includes a length of the aligned region, and the lengths of the aligned regions are compared to the pixels in the first eye image and the second eye image. The difference in the data is determined. The length of each of the first to fourth mask regions further includes a length of a first extended frame region.

According to another exemplary embodiment, a method for processing a three-dimensional image is provided, including: generating a first eye image and a second eye image according to one of the original two-dimensional images; a first mask region and a second mask region are respectively generated at a first boundary and a second boundary; and a first boundary and a second boundary are respectively generated at the first boundary of the second eye a third mask area and a fourth mask area. The lengths of the first to fourth mask regions are respectively the first to fourth lengths, and the first to fourth lengths are not zero, and the first length is not equal to the third length, and the second length is not Equal to the fourth length.

In order to better understand the above and other aspects of the present invention, the following specific embodiments, together with the drawings, are described in detail below:

Please refer to FIG. 1 , which shows a flow chart of a three-dimensional image processing method according to an embodiment. In step 110, the first eye image and the second eye image of the three-dimensional image are generated according to the original two-dimensional image. The first eye picture is, for example but not limited to, one of a left eye picture and a right eye picture, and the second eye picture is the other of the left eye picture and the right eye picture. For example, in step 110, the original two-dimensional image is translated by a translation amount in opposite directions to generate a first eye image and a second eye image, respectively.

In step 120, the difference between the pixel data in the first eye image and the second eye image is compared to determine the length of an aligned region.

In step 130, a first mask region and a second mask region are respectively generated at a first boundary of the first eye image and a second boundary according to the length of the comparison region, and A third mask region and a fourth mask region are respectively generated at one of the first boundary and the second boundary of the second eye.

In step 140, a first extended bezel area is further extended from each of the first to fourth mask regions.

Optionally, in step 150, each of the second and third mask regions is extended beyond a second extended frame region. It is worth noting that in Figure 1, step 150 is framed by a dashed line, representing that this step is an optional step that can be implemented depending on design requirements. In addition, in another embodiment, step 140 can also be used as an optional step, but step 150 is implemented.

The embodiments shown in Figs. 2A to 3B will be used hereinafter to detail the details of steps 120 to 150 in the three-dimensional image processing method shown in Fig. 1. In the 2A to 3B drawings, like reference symbols represent similar meanings. In addition, in the embodiments shown in FIGS. 2A to 3B, the length relationship between the first to fourth mask regions and the comparison region in steps 120 to 150, and the first extended frame region and The length relationship between the second extended frame regions.

Remote image processing

2A is a diagram showing image processing of the left boundary LB of the left eye picture and the left boundary LB of the right eye picture of the remote 3D image according to an embodiment. FIG. 2B is a schematic diagram showing image processing of the right border RB of the left eye picture and the right border RB of the right eye picture according to the remote three-dimensional image according to the embodiment. When the human eye views the distant three-dimensional image of FIG. 2A and FIG. 2B, the human eye feels that the three-dimensional image is displayed far away. In other words, the human eye feels that the 3D image is displayed behind the screen.

Please refer to both Figure 1 and Figure 2A. 2D represents the original 2D image, and LF and RF represent the left eye picture and the right eye picture, respectively. LB and RB represent a left boundary LB and a right boundary RB. The visible area represents the area that the human eye can view when viewing a 2D image or a 3D image.

First, step 110 of Fig. 1 will be described. In FIG. 2A, the left-to-right pixels of a certain pixel of the two-dimensional image 2D are listed as A, B, C, D, E, F, . Therefore, in step 110, the two-dimensional image 2D can be shifted to the left by a shift amount to generate the left-eye image LF, and the two-dimensional image 2D is shifted to the right by the shift amount to generate the right-eye image RF. It should be noted that the actual resolution of the two-dimensional image is not limited to the example description of the embodiment. Further, the shift amount is exemplified as 4 pixels, but it is to be understood that the present invention is not limited thereto. For example, the amount of translation may be 1/2 pixel, 1/4 pixel, 1/8 pixel, or the like.

Next, step 120 of Fig. 1 will be described. It can be observed that in FIG. 2A, since the right-eye picture RF is shifted to the right by 4 pixels, the four pixel values of the left boundary LB of the right-eye picture RF have been removed without meaning (here X1~) X4 indicates it). On the other hand, since the left-eye picture LF is shifted to the left by 4 pixels, the original leftmost 4 pixel values A to D of the left-eye picture LF are removed from the visible area and are not visible.

Therefore, after comparing the left-eye picture LF with the right-eye picture RF, it can be found that at the left boundary LB, the pixel data X1~X4 and A~D appear on the right-eye picture RF but not on the left-eye picture LF. . Therefore, the region where the pixel data X1~X4 and A~D are located can be taken as a comparison area M1, and the length is equal to twice the translation amount.

Next, step 130 of Fig. 1 is illustrated, which corresponds to step 210 of Fig. 2A. In step 210, a mask area LF_ML is generated on the left boundary LB of the left-eye picture LF according to the length of the comparison area M1, and a mask area RF_ML is generated on the left boundary LB of the right-eye picture RF. The length of the mask area LF_ML at the left boundary LB of the left eye picture is temporarily 0. The mask area RF_ML of step 210 includes the comparison area M1, or the length of the mask area RF_ML includes the length Lcom of the comparison area M1. Therefore, after step 210, the left boundary LB of the left-eye picture LF is masked by 0 pixels, and the left boundary LB of the right-eye picture RF is masked by Lcom pixels.

In short, steps 120 and 130 compare the left-eye picture LF with the right-eye picture RF, and take the area of the pixel data that does not appear on the left-eye picture LF as the comparison area M1, and add it. Mask. The principle of steps 120 and 130 is that, in the eyes of the human eye, the left and right eyes must see the same pixel, and the human eye can have a way to focus, that is, if only one eye sees one pixel and the other eye does not see In this case, the human eye cannot focus on this pixel. In the case where the alignment mask is not set, the pixels A to D are only in the right-eye picture RF and not in the left-eye picture LF, so the human eye cannot focus on the pixels A to D. Therefore, in the present embodiment, the alignment area is masked, and the point where the human eye cannot see the focus can be avoided, so that the viewing comfort can be improved.

Next, step 140 of Fig. 1 is illustrated, which corresponds to step 220 of Fig. 2A. In step 220, the first extended frame area n1 is extended from the mask area LF_ML of the left-eye picture LF and the mask area RF_ML of the right-eye picture RF. That is, the length of the mask area LF_ML of the left-eye picture LF and the mask area RF_ML of the right-eye picture RF all include the length Lvf of the first extended frame area n1. That is, in step 220, the left boundary LB of the left-eye picture LF is further masked by Lvf pixels, and the left boundary LB of the right-eye picture RF is further masked by Lvf pixels. In this embodiment, for example, the pixels E and F of the left-eye picture LF and the pixels E and F of the right-eye picture RF are blocked.

After step 220 is performed, the length of the mask area LF_ML of the left-eye picture LF is Lvf, and the mask area RF_ML of the right-eye picture RF is Lcom+Lvf. The principle of step 220 is that if the human eye views the left-eye picture LF and the right-eye picture RF of step 220 of FIG. 2A, the human eye feels that the frame and the image are in the same visual plane, and the human eye can focus on the first extended frame. Area n1.

Next, step 150 of Fig. 1 is illustrated, which corresponds to step 230 of Fig. 2A. In step 230, the second extended frame area k1 is further extended from the mask area LF_ML of the left-eye picture LF, but the mask area RF_ML of the right-eye picture RF does not extend the second extended frame area k1. That is, in step 230, the left border LB of the left-eye picture LF is further masked by Lfs pixels. Therefore, after the step 230 is performed, the length of the mask area LF_ML of the left-eye picture LF is Lvf+Lfs, and the mask area RF_ML of the right-eye picture RF is Lcom+Lvf.

In step 230, the virtual border formed by the mask area and the three-dimensional image are in different visual planes, that is, the human eye views the virtual border as if it were a photo frame. For example, the human eye may feel that the three-dimensional image is concave in the virtual border. In this way, the comfort of the human eye to view the three-dimensional image can be improved. If the mask area RF_ML of the right-eye picture RF also includes the second extended frame area k1, the virtual black frame and the three-dimensional image will be in the same visual plane, which cannot improve the comfort of the human eye. The length of the second extended bezel area k1 is Lfs. It should be noted that in other possible embodiments, it is also possible to make the three-dimensional image protrude from the virtual frame when viewed, which is also within the spirit of the present invention.

Next, please refer to both Figure 1 and Figure 2B. First, step 120 of Fig. 1 will be described. It can be observed that in FIG. 2B, since the left-eye picture LF is shifted to the left by 4 pixels, the four pixel values of the right boundary RB of the left-eye picture LF have been removed without meaning (here Y1~) Y4 indicates). On the other hand, since the right-eye picture RF is shifted to the right by 4 pixels, the original rightmost 4 pixel values A1 to D1 of the right-eye picture RF are removed from the visible area and are not visible.

Therefore, after comparing the left-eye picture LF with the right-eye picture RF, it can be found that at the right boundary RB, the pixels A1, B1, C1, D1, Y1, Y2, Y3, and Y4 appear on the left-eye picture LF but are not Appears in the right eye screen RF. Therefore, the region where the pixel data Y1, Y2, Y3, Y4, A1, B1, C1, and D1 are located can be taken as a comparison region M2, and the length thereof is equal to twice the amount of translation.

Next, step 130 of Fig. 1 is illustrated, which corresponds to step 240 of Fig. 2B. In step 240, a mask area LF_MR is generated on the right border RB of the left-eye picture LF according to the length of the comparison area M2, and a mask area RF_MR is generated on the right border RB of the right-eye picture RF. The mask area LF_MR includes the comparison area M2, which is also Lcom in length. That is, in step 240, the right border RB of the left-eye picture LF is masked by Lcom pixels, and the right border RB of the right-eye picture RF is masked by 0 pixels.

Next, step 140 of FIG. 1 is illustrated, which corresponds to step 250 of FIG. 2B. In step 250, the first extended frame region n2 is extended from the mask region LF_MR of the left-eye picture LF and the mask region RF_MR of the right-eye frame RF. That is, the length of the mask area LF_MR of the left-eye picture LF and the mask area RF_MR of the right-eye picture RF further include the length Lvf of the first extended frame area n2. The lengths of the first extended frame regions n1 and n2 are both Lvf.

That is, in step 250, the right border RB of the left-eye picture LF is further masked by Lvf pixels, and the right border RB of the right-eye picture RF is further masked by Lvf pixels. Therefore, after the step 250 is performed, the length of the mask area LF_MR of the left-eye picture LF is Lcom+Lvf, and the mask area RF_MR of the right-eye picture RF is Lvf. When the human eye views the left eye picture LF and the right eye picture RF of step 250 of FIG. 2B, the human eye will think that the frame and the image are in the same visual plane.

Next, step 150 of FIG. 1 is illustrated, which corresponds to step 260 of FIG. 2B. In step 260, the second extended frame area k2 is further extended from the mask area RF_MR of the right eye picture RF, but the mask area LF_MR of the left eye picture LF does not extend out of the second extended frame area k2 (step 150). The reason is the same as step 230 of FIG. 2A. The length of the second extended bezel area k2 is also Lfs. That is, in step 260, the right border RB of the right eye picture LF is further masked by Lfs pixels. After the step 260 is performed, the length of the mask area LF_MR of the left-eye picture LF is Lcom+Lvf, and the mask area RF_ML of the right-eye picture RF is Lvf+Lfs.

As can be seen from FIGS. 2A and 2B, for the left-eye picture LF, the mask area LF_ML of its left border LB is asymmetrical with the mask area LF_MR of its right border RB. Similarly, for the right-eye picture RF, the mask area RF_ML of its left border LB and its mask area RF_MR of the right border RB are also asymmetrical.

Near-field image processing

Please refer to Figures 3A and 3B. FIG. 3A is a diagram showing image processing of the left boundary LB of the left-eye picture and the left boundary LB of the right-eye picture of the near-side three-dimensional image according to an embodiment. FIG. 3B is a view showing the image processing of the right border RB of the left eye picture and the right border RB of the right eye picture according to the near-side three-dimensional image according to the present embodiment. When the human eye views the 3D images of the 3A and 3B, the human eye feels that the 3D image is displayed closer. In other words, the human eye feels that the 3D image is displayed on the front of the screen.

Please refer to both Figure 1 and Figure 3A. First, step 120 of Fig. 1 will be described. It can be observed that in FIG. 3A, since the left-eye picture LF' is shifted to the right by 4 pixels, the 4 pixel values of the left boundary LB of the left-eye picture LF' have been removed and are not visible (here X1'~X4' indicates it). On the other hand, since the right-eye picture RF' is shifted to the left by 4 pixels, the original leftmost 4 pixel values A' to D' of the right-eye picture LF' are removed from the visible area and are not visible.

Therefore, after comparing the left-eye picture LF' with the right-eye picture RF', it can be found that at the left boundary LB', the pixel data X1'~X4' and A'~D' appear on the left-eye picture LF', but It does not appear in the right eye picture RF'. Therefore, the region where the pixel data X1'~X4' and A'~D' are located can be taken as a matching region M1', and the length thereof is equal to twice the amount of translation.

Next, step 130 of FIG. 1 is illustrated, which corresponds to step 310 of FIG. 3A. In step 310, according to the length of the comparison area M1', a mask area LF_ML' is generated at the left boundary LB' of the left-eye picture LF', and a mask area RF_ML' is generated at the left boundary LB' of the right-eye picture RF'. . The length of the mask area RF_ML' of step 310 is temporarily zero. The mask area LF_ML' of step 310 includes the alignment area M1', or the length of the mask area LF_ML' includes the length Lcom' of the comparison area M1'. That is, in step 310, the left boundary LB' of the left-eye picture LF' is masked by Lcom' pixels, and the left boundary LB of the right-eye picture RF' is masked by 0 pixels.

Next, step 140 of FIG. 1 is illustrated, which corresponds to step 320 of FIG. 3A. In step 320, the first extended bezel area n1' is extended from the mask area LF_ML' of the left-eye picture LF' and the mask area RF_ML' of the right-eye picture RF' (step 140). The length of the mask area LF_ML' of the left-eye picture LF' and the mask area RF_ML' of the right-eye picture RF' include the length Lvf' of the first extended frame area n1'.

In other words, in step 320, the left boundary LB' of the left-eye picture LF' is further masked by the Lvf' pixels, and the left boundary LB' of the right-eye picture RF' is further masked by the Lvf' picture. Prime. Therefore, after the step 320 is performed, the length of the mask area LF_ML' of the left-eye picture LF' is Lcom'+Lvf', and the mask area RF_ML' of the right-eye picture RF' is Lvf'. When the human eye views the left eye picture LF' and the right eye picture RF' of step 320 of Fig. 3A, the human eye will feel that the frame and the image are in the same visual plane.

Next, step 150 of FIG. 1 is illustrated, which corresponds to step 330 of FIG. 3A. In step 330, the second extended frame area k1' is extended from the mask area LF_ML' of the left-eye picture LF', but the mask area RF_ML' of the right-eye picture RF' does not extend out of the second extended frame area. K1' (step 150), the reason is as in step 230. That is, in step 330, the left boundary LB' of the left-eye picture LF' is further masked by Lfs' pixels. Therefore, after the step 330 is performed, the length of the mask area LF_ML' of the left-eye picture LF' is Lcom'+Lvf'+Lfs', and the mask area RF_ML' of the right-eye picture RF' is Lvf'.

Please refer to both Figure 1 and Figure 3B. First, step 120 of Fig. 1 will be described. It can be observed that in FIG. 3B, since the right eye picture RF' is shifted to the right by 4 pixels, the 4 pixel values of the right border RB' of the right eye picture RF' have been removed (here Y1) '~Y4' indicates). On the other hand, since the left-eye picture LF' is shifted to the right by 4 pixels, the original rightmost 4 pixel values A1' to D1' of the left-eye picture LF' are removed from the visible area and are not visible.

Therefore, after comparing the left-eye picture LF' and the right-eye picture RF', it can be found that in the 3B picture, at the right boundary RB', the pixel data Y1'~Y4' and A1'~D1' appear on the right. The eye picture RF', but does not appear in the left eye picture LF', so the area where the pixel data Y1'~Y4' and A1'~D1' are located can be taken as a comparison area M2'.

Next, step 130 of Fig. 1 is illustrated, which corresponds to step 340 of Fig. 3B. In step 340, a mask area LF_MR' is generated at the right border RB' of the left-eye picture LF' according to the length of the comparison area M2', and a mask area RF_MR' is generated at the right border RB' of the right-eye picture RF'. . The mask area RF_MR' includes the alignment area M2', which is also Lcom' in length. The length of the mask area LF_ML' of step 340 is temporarily zero. That is, in step 340, the right border RB' of the right-eye picture RF' is masked by Lcom' pixels, and the right border RB' of the left-eye picture LF' is masked by 0 pixels.

Next, step 140 of FIG. 1 is illustrated, which corresponds to step 350 of FIG. 3B. In step 350, the first extended frame region n2' extends from the mask region LF_MR' of the left-eye picture LF' and the mask region RF_MR' of the right-eye picture RF'. That is, the length of the mask area LF_MR' of the left-eye picture LF' and the mask area RF_MR' of the right-eye picture RF' further include the length Lvf' of the first extended frame area n2'. The lengths of the first extended frame regions n1' and n2' are both Lvf'.

That is, in step 350, the right border RB' of the left-eye picture LF' is further masked by Lvf' pixels, and the right border RB' of the right-eye picture RF' is further masked by Lvf' Picture. Therefore, after the step 350 is performed, the length of the mask area LF_MR' of the left-eye picture LF' is Lvf', and the mask area RF_MR' of the right-eye picture RF' is Lcom'+Lvf'. When the human eye views the left eye picture LF' and the right eye picture RF' of step 350 of Fig. 3B, the human eye will feel that the frame and the image are in the same visual plane.

Next, step 150 of FIG. 1 is illustrated, which corresponds to step 360 of FIG. 3B. In step 360, the second extended frame area k2' is extended from the mask area RF_MR' of the right eye picture RF', but the mask area LF_MR' of the left eye picture LF' does not extend out of the second extended frame area. K2', the reason is the same as step 230 of FIG. 2A. The length of the second extended bezel area k2' is also Lfs'. That is, in step 360, the right border RB' of the right eye picture LF' is further masked by Lfs' pixels. Therefore, after the step 360 is performed, the length of the mask area LF_MR' of the left-eye picture LF' is Lvf', and the mask area RF_ML' of the right-eye picture RF' is Lcom'+Lvf'+Lfs'.

As can be seen from Figs. 3A and 3B, for the left-eye picture LF', the mask area LF_ML' of its left boundary LB' is asymmetrical with the mask area LF_MR' of its right border RB'. Similarly, for the right-eye picture RF', the mask area RF_ML' of its left border LB' and its mask area RF_MR' of the right border RB' are also asymmetrical.

In short, in the above embodiment, the left boundary mask region and the right border mask region of the last generated left-eye picture have a first length and a second length, respectively, and the left boundary of the generated right-eye picture is finally covered. If the cover region and the right boundary mask region respectively have a third length and a fourth length, then the first to fourth lengths are not 0, and the first length is not equal to the third length, and the second length is not equal to the fourth length. Still further, the first length and the fourth length are substantially equal, and the second length and the third length are substantially equal.

Additionally, the first length can be greater than the third length, and the fourth length can be greater than the second length. Additionally, the first length and the fourth length are substantially equal, and the second length is substantially equal to the third length.

In an example, the first length and the fourth length may both be equal to Lcom+Lvf, and the second and third lengths may be equal to Lvf, where Lcom represents the length of the comparison area, which is the first eye picture and the The second eye image is compared to the length of one of the aligned regions occupied by only one of the pixel data, such as the first eye image or the second eye image relative to the original two-dimensional image. One of the screens is twice the length of the translation. Lvf stands for virtual border length, which can be set according to design requirements.

In another example, the first length and the fourth length may both be equal to Lcom+Lvf, and the second and third lengths may both be equal to Lvf+Lfs, wherein Lcom represents the length of the comparison region, which may be based on the foregoing Compare the way or get. In addition, Lvf represents the length of the virtual border, and Lfs represents the length of the frame shift, both of which can be set according to design requirements.

In another example, the first length and the fourth length may both be equal to Lcom+Lvf+Lfs, and the second and third lengths are both equal to Lvf, wherein Lcom represents the length of the comparison region, and Lvf represents the virtual frame length. Lfs represents the length of the frame translation, and the three are set according to the above methods.

Further, in the present embodiment, for each pixel column, the amount of translation and the length of the alignment area thereof may be identical to each other or different from each other. What is more, the amount of translation corresponding to each pixel column and the length of the alignment area can be changed according to the number of columns of the pixel column. For example, the translation amount of the top pixel column and the length of the comparison region can be larger, and the translation amount of the bottom pixel column and the length of the comparison region can be smaller, which can enhance the human eye to view the three-dimensional image. Comfort.

In addition, in the above embodiment, since the virtual frames on both sides of the left-eye picture are asymmetrical to each other, the content of the original two-dimensional image can be retained as much as possible. In addition, in the above embodiment, the virtual frame can be filled with black pixels or white pixels (that is, the virtual frame can be a black frame or a white frame), which is also within the spirit of the present invention.

In summary, although the present invention has been disclosed above by way of example, it is not intended to limit the present invention. Those who have ordinary knowledge in the technical field of the present invention can make various changes and refinements without departing from the spirit and scope of the present case. Therefore, the scope of protection of this case is subject to the definition of the scope of the patent application attached.

110~130, 210~260, 310~360. . . step

Fig. 1 is a flow chart showing a method of processing a three-dimensional image according to an embodiment of the present invention.

Fig. 2A is a view showing the image processing of the left boundary LB of the left-eye picture and the left boundary LB of the right-eye picture of the remote three-dimensional image according to the present embodiment.

FIG. 2B is a schematic diagram showing image processing of the right border RB of the left eye picture and the right border RB of the right eye picture according to the remote three-dimensional image according to the embodiment.

Fig. 3A is a view showing the image processing of the left boundary LB of the left-eye picture and the left boundary LB of the right-eye picture of the near-side three-dimensional image according to the present embodiment.

FIG. 3B is a view showing the image processing of the right border RB of the left eye picture and the right border RB of the right eye picture according to the near-side three-dimensional image according to the present embodiment.

110~150. . . step

Claims (18)

A method for processing a three-dimensional image, comprising: generating a first eye image and a second eye image according to one of the original two-dimensional images; at a first boundary of the first eye image a first mask area and a second mask area are respectively generated at the two boundaries; and a third mask area and a fourth are respectively generated at a first boundary and a second boundary of the second eye picture a mask area; wherein a length of each of the first and the fourth mask areas includes a length of the comparison area, the length of the comparison area being compared to the first eye and the second eye The difference between the pixel data in the picture is determined; and the length of each of the first to fourth mask regions further includes the length of a first extended frame region. The three-dimensional image processing method of claim 1, wherein: the comparison region of the first mask region comprises pixel data that is not present in the second eye image by the comparison; The aligned region of the second mask region includes pixel data that is not present in the first eye frame through the alignment. The method of claim 3, wherein the comparison region of the first mask region is located at the first boundary of the first eye frame and does not appear in the second eye frame. The pixel data in the middle, and the comparison region of the fourth mask region includes pixel data located at the second boundary of the second eye image and not present in the first eye image. The method of claim 3, wherein the step of generating the first eye image and the second eye image according to the original two-dimensional image comprises: Translating a translation amount in opposite directions to generate the first eye picture and the second eye picture, respectively. The method of claim 3, wherein the length of the aligned area of each of the first and fourth mask regions is equal to twice the amount of the translation. The three-dimensional image processing method of claim 1, wherein the length of the first extended frame region of each of the first to fourth mask regions is equal. The method of claim 3, wherein the length of each of the second and third mask regions further comprises a length of a second extended frame region. The three-dimensional image processing method of claim 7, wherein the length of the second extended frame region of each of the second and third mask regions is equal. A method for processing a three-dimensional image, comprising: generating a first eye image and a second eye image according to one of the original two-dimensional images; at a first boundary of the first eye image a first mask area and a second mask area are respectively generated at the two boundaries; and a third mask area and a fourth are respectively generated at a first boundary and a second boundary of the second eye picture a mask area; wherein the lengths of the first to fourth mask areas are respectively first to fourth lengths, the first to fourth lengths are not zero, and the first length is not equal to the third length, The second length is not equal to the fourth length. The method of claim 3, wherein the first length is greater than the third length, and the fourth length is greater than the second length. The three-dimensional image processing method of claim 9, wherein the first length and the fourth length are substantially equal, and the second length is substantially equal to the third length. The method of claim 3, wherein the first length and the fourth length are both equal to Lcom+Lvf, and the second and third lengths are equal to Lvf, wherein Lcom represents an aligned region. Length, Lvf represents a virtual border length. The method of claim 3, wherein the first length and the fourth length are both equal to Lcom+Lvf, and the second and third lengths are equal to Lvf+Lfs, wherein Lcom represents a ratio For the length of the region, Lvf represents a virtual border length, and Lfs represents a border translation length. The method of claim 3, wherein the first length and the fourth length are both equal to Lcom+Lvf+Lfs, and the second and third lengths are equal to Lvf, wherein Lcom represents a ratio For the length of the region, Lvf represents a virtual border length, and Lfs represents a border translation length. The method of claim 3, wherein the length of the comparison region is such that the first eye image and the second eye image are compared and only one of the pixel data is present. One of the lengths of the comparison area. The method of claim 3, wherein the length of the comparison region is twice the translation length of the first eye image or the second eye image relative to the original two-dimensional image. The method of processing a three-dimensional image according to claim 14, wherein the length of the comparison region is such that the first eye image and the second eye image are compared and only one of the pixel data is present. One of the lengths of the comparison area. The method of processing a three-dimensional image according to claim 17, wherein the length of the comparison region is substantially equal to a translation length of the first eye image or the second eye image relative to the original two-dimensional image. 2 times.